The post-crisis Irish voter: Voting behaviour in the Irish 2016 general election 9781526122650

Provides the definitive study of voting behaviour in the 2016 Irish election

126 108 6MB

English Pages [288] Year 2018

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Front matter
Contents
List of figures and tables
Notes on contributors
Editors’ preface
Ireland’s post-crisis election
Mining the ballot: Preferences and transfers in the 2016 election
Ideological dimensions in the 2016 elections
Social and ideological bases of voting
Party identification in the wake of the crisis: A nascent realignment?
Why did the ‘recovery’ fail to return the government?
Party or candidate?
Political fragmentation on the march: Campaign effects in 2016
The impact of gender quotas on voting behaviour in 2016
What do Irish voters want from and think of their politicians?
Popularity and performance? Leader effects in the 2016 election
Appendix The INES 2016 questionnaires
Index
Recommend Papers

The post-crisis Irish voter: Voting behaviour in the Irish 2016 general election
 9781526122650

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

The posT-crisis irish voTer Vot ing b e haVio u r i n t he irish 20 1 6 ge ne ral e l ect i o n

EDITED BY MIchaEl Marsh, DavID M. FarrEll anD ThErEsa rEIDY

The post-crisis Irish voter

The post-crisis Irish voter Voting behaviour in the Irish 2016 general election

Edited by Michael Marsh, David M. Farrell and Theresa Reidy

MANCHESTER UNIVERSITY PRESS

Copyright © Manchester University Press 2018 While copyright in the volume as a whole is vested in Manchester University Press, copyright in individual chapters belongs to their respective authors, and no chapter may be reproduced wholly or in part without the express permission in writing of both author and publisher. Published by Manchester University Press Altrincham Street, Manchester M1 7JA www.manchesteruniversitypress.co.uk British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data applied for ISBN 978 1 5261 2264 3 hardback First published 2018 The publisher has no responsibility for the persistence or accuracy of URLs for any external or third-party internet websites referred to in this book, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Typeset by Deanta Global Publishing Services, Dublin, Ireland

Contents

List of figures and tables page vii Notes on contributors xiii Editors’ preface xvi 1

Ireland’s post-crisis election Michael Marsh, David M. Farrell and Theresa Reidy

1

2

Mining the ballot: Preferences and transfers in the 2016 election Kevin Cunningham

14

3

Ideological dimensions in the 2016 elections Kevin Cunningham and Johan A. Elkink

33

4

Social and ideological bases of voting John Garry

63

5

Party identification in the wake of the crisis: A nascent realignment? Rory Costello

6

Why did the ‘recovery’ fail to return the government? Michael Marsh

7

Party or candidate? Michael Courtney and Liam Weeks

126

8

Political fragmentation on the march: Campaign effects in 2016 Theresa Reidy and Jane Suiter

145

9

The impact of gender quotas on voting behaviour in 2016 Gail McElroy

165

82 99

vi

Contents

10

What do Irish voters want from and think of their politicians? David M. Farrell, Michael Gallagher and David Barrett

190

11

Popularity and performance? Leader effects in the 2016 election Stephen Quinlan and Eoin O’Malley

209

Appendix: The INES 2016 questionnaires Index

233 266

List of figures and tables

Figures 1.1

Party support levels during January and February 2016 and election result

2.1

First preference vote shares since 1922

16

2.2

The estimated relationship between the quota received and the probability of winning a seat

29

3.1a

Mean support for Fine Gael and predicted support for Fine Gael within attitudinal categories

41

3.1b Mean support for Fianna Fáil and predicted support for Fianna Fáil within attitudinal categories 3.1c

Mean support for Sinn Féin and predicted support for Sinn Féin within attitudinal categories

2

41 42

3.1d Mean support for Labour and predicted support for Labour within attitudinal categories

42

3.2a and 3.2b  Irish and UK voter’s self-placement and placement of political parties

47

3.3

Effect plots explaining the effect of [1] voter-party proximity and of [2] party extremism on vote choice

50

3.4

Graphical representation of the latent dimensions estimated by the multidimensional scaling algorithm, summarizing distances between parties based on the highest preference for each party assigned by each respondent

53

viii

List of figures and tables

4.1

Expectations of the social and ideological bases of vote choice

66

4.2

Wording of the ideology questions

67

4.3

Voters’ ideological beliefs

69

4.4

Extent to which ideological views can be explained by socio-demographic characteristics

71

4.5

Party voters by age, class and education

73

4.6

Party voters by Catholic, religiosity and residence

74

4.7

The ideological profile of each party’s voters

76

5.1

Levels of partisanship (INES data), 2002–2016

90

5.2

Levels of partisanship, 2002–2016, by generation

91

5.3

Direction of partisanship 2002–2016

93

6.1

Economic indicators 2011–2016

102

6.2

Regional levels of unemployment 2007–2016

104

6.3

Consumer sentiment 2011–2016

105

6.4

Respondents’ estimates of unemployment in 2016

109

6.5

Differences in most positive sociotropic evaluations by class and gender based on all INES polls 2016

112

6.6

Sociotropic effects by attribution and vote in 2011 based on INES3 115

6.7

Impact of sociotropic evaluations on Labour share of government vote based on all INES polls 2016

116

6.8

Comparison of sociotropic and pocketbook voting effects based on all INES polls 2016

117

7.1

Proportions of party- and candidate-centred voters casting sequential party ballots by year and party

136

8.1

Party support levels during January and February 2016 and election result

151

8.2

Timing of vote decision 1997–2016

154

9.1

Proportion of women running and elected in Irish elections 1948–2016

167

9.2

Proportion of women running for the four main parties 2002–2016 168

List of figures and tables

ix

9.3

Odds of women winning 1973–2011

169

9.4a

Support for gender quota among general election candidates

171

9.4b Support for gender quota among voters

171

9.5

Impact of gender on candidate success 2016

173

9.6

Impact of candidate gender and education on success 2016.

174

9.7a

Factors affecting voting for female candidates (any preference)

177

9.7b Factors affecting voting for female candidates (high preference)

177

9.8

Irish voter ‘Big Five’ trait distribution

180

9.9

Impact of ‘Big Five’ traits on casting a first preference for a female candidate.

181

11.1 Leader popularity in 2011 and 2016 Irish general elections (%)

215

11.2 Average predicted effect of voting for the four main parties dependent on the leader’s performance in the television debates

221

11.3 Leadership and party popularity and vote choice: 2016 Irish general election (%)

223

11.4 Average predicted effect of voting for Fine Gael, dependent on feelings towards Enda Kenny

224

11.5 Average predicted effect of voting for Fianna Fáil, dependent on feelings towards Micheál Martin

225

11.6 Average predicted effect of voting for Sinn Féin, dependent on feelings towards Gerry Adams

226

11.7 Average predicted effect of voting for Labour, dependent on feelings towards Joan Burton

227

Tables 1.1

Results of the 2016 general election

2.1

Voting and non-voting in 2016, by voting and non-voting in 2011

17

2.2

The proportion of voters switching from their party of choice, staying with their party and not voting from one election cycle to the next, 2011–2016 and 1997–2002 (per cent)

18

Number of different parties voted for by number of options (per cent) and first preference choice

20

2.3

3

x

List of figures and tables

2.4

Distribution of successive party preferences excluding cases where the parties did not stand

22

2.5

Distribution of first and second preferences of voters with an option to give a second preference to the same party

24

2.6

Patterns of party choice: Proportion of first preferences for each party going to each possible second next preferred party

25

2.7

Party first preference vote shares, percentage of votes transferred from each party and their overall contribution to the share of all transferred votes as taken from the results of the 2016 election (2016 Electoral Data)

27

Number of seats won, number of occasions the party would have won without any additional counting, modelled estimates of expected seats and effective seat bonus due to vote management and transfers

30

Regression analysis estimating the extent to which a respondent likes Fine Gael, Fianna Fáil, Sinn Féin, Labour, and the difference in the extent to which they like Fianna Fáil over Fine Gael

44

3.2

Regression analysis of estimated left–right position

48

3.3

Average assessments of supporters, non-supporters and the average position of supporters themselves

49

3.4

Party second preference share of party first preferences

52

4.1

Social bases of ideological beliefs

70

4.2

Social bases of ideological beliefs: Binary logistic regression models 71

4.3

Social and ideological bases of vote choice: Binary logistic regression models

77

4.4

Vote dynamics 2011 to 2016: Profiling FG stickers and switchers

78

5.1

Stability of vote preference by party attachment, 2002–2007

86

5.2

Perceived changes in conditions between 2002 and 2007 (respondents grouped according to 2002 partisanship)

87

5.3

Direction of partisanship 2016, by generation

94

6.1

Regional variations in unemployment and employment (figures are percentages)

103

6.2

Retrospective perceptions of the country as a whole

106

2.8

3.1

List of figures and tables

xi

6.3

Retrospective perceptions of the household finances

107

6.4

Sociotropic evaluations by unemployment and employment levels in region of respondent (ologit estimations)

111

7.1

Reasons for voting: Party or candidate

130

7.2

Would vote for the same candidate if he or she stood for a different party

131

7.3

Candidate/party index from close-ended questions

133

7.4

Patterns of sequential party voting by candidate-centredness of the first preference vote (per cent), 2002–2016

135

7.5

Patterns of sequential party voting by candidate-centredness of the first preference vote (per cent), 2002–2016

136

7.6

Independent voters

140

8.1

Demographic profile of election day deciders, campaign deciders and early deciders

155

8.2

Timing of vote choice by vote decision

156

8.3

Timing of vote choice by party

158

8.4

Timing of vote choice and most important issue at the election

160

8.5

Timing of vote choice by national or local policy considerations

160

8.6

Regression: Timing of vote choice

160

10.1 Importance of parliamentary representatives’ possession of certain characteristics

192

10.2 Importance of various aspects of the role of parliamentary representative 195 10.3 Impact of local versus cosmopolitan orientation on perceptions of role of parliamentary representative

196

10.4 Extent of contact between citizens and TDs, 2002–2016

197

10.5 Contact and political engagement: Percentage who have contacted a TD in preceding period 2002–2016

198

10.6 Modes of contact with TDs in 2002 and 2016

199

10.7 On what subject were TDs contacted, 2011 and 2016

201

10.8 What do Irish voters think of their politicians?

203

xii

List of figures and tables

11.1 Average popularity of leaders among voters in the 2016 general election and change in mean rating since the 2011 election

216

11.2 Assessments of leader performance in television debates among all voters, partisans and non-partisans and impact on vote choice in 2016 Irish general election (%)

219

11.3 Proportion of voters voting for each party dependent on leader popularity 222

Appendix tables A3.1 Modelling the relationship between a respondent’s party choice and the difference between where they put themselves and where they put each party on the left–right scale

58

A6.1 Regression analysis of sociotropic evaluations on vote … underlying Figure 6.6

121

A6.2 Regression analysis of sociotropic evaluations on Labour share of coalition vote … underlying Figure 6.7

122

A6.3 Regression analysis of sociotropic and pocketbook evaluations on coalition vote … underlying Figure 6.8

123

A7.1 Data for Figure 7.1

142

A9.1 Aggregate analysis 2016

183

A9.2 Casting a (high) preference for a female candidate, 2016

184

A9.3 Big Five questions

185

Notes on contributors

David Barrett is Researcher at the Geary Institute in University College Dublin. He undertook his PhD research at Trinity College Dublin, where he examined infighting within political parties. He has a long-standing interest in political parties, electoral behaviour and public opinion and has written on Irish and Greek politics. Rory Costello is Lecturer in Politics at the University of Limerick. His research focuses on EU politics, Irish politics, political parties and democratic representation. He has (co)-authored articles in journals such as Party Politics, West European Politics, Electoral Studies, Journal of European Public Policy, European Union Politics and American Journal of Political Science. Michael Courtney is Postdoctoral Researcher in the School of Law and Government at Dublin City University. He currently works on VOX-POL, a European Union Framework Programme 7 (FP7)-funded network of excellence, studying violent online political extremism. He has published on political parties in the Journal of Legislative Studies and Irish Political Studies, and will release a co-authored book on political communications in Irish elections with Manchester University Press in 2018 with Michael Breen, Iain McMenamin, Eoin O’Malley and Kevin Rafter. Kevin Cunningham is Lecturer in Statistics at Dublin Institute of Technology. He has published a number of book chapters and peer-reviewed journal articles on public policy, political parties, elections and campaigns. Kevin has worked for a number of political parties in Ireland, the United Kingdom, Australia and France. He runs his own research company and his work has been published in all major Irish newspapers and broadcasters. Johan A. Elkink is Associate Professor in Politics at the School of Politics and International Relations at University College Dublin. He specializes in quantitative methods in political science, particularly spatial econometrics

xiv

Notes on contributors

and their applications in democratization and voting behaviour. He has coauthored reports on voting behaviour in the Irish referendums on the Lisbon Treaty and co-edited The Act of Voting (Routledge). His work has appeared in the Journal of Politics, Comparative Political Studies, European Journal of Political Research and Electoral Studies. David M. Farrell, MRIA, is Head of Politics and International Relations at University College Dublin. A specialist in the study of representation, elections and parties, his most recent books include Political Parties and Democratic Linkage (co-authored, Oxford University Press, 2011) and A Conservative Revolution? Electoral Change in Twenty-First Century Ireland (co-edited, Oxford University Press, 2017). His current work is focused on constitutional deliberation, and in that capacity he is the research leader of the ongoing Irish Citizens’ Assembly. Michael Gallagher, MRIA, is Head of the Department of Political Science and Professor of Comparative Politics at Trinity College, University of Dublin. He is co-editor of  Politics in the Republic of Ireland  (6th edn, Abingdon, 2018),  Representative Government in Modern Europe  (5th edn, New York, 2011),  The Politics of Electoral Systems  (Oxford, 2008),  Days of Blue Loyalty: The Politics of Membership of the Fine Gael Party (Dublin, 2002) and a number of books in the How Ireland Voted series. In 1989, he devised the least squares index, the standard measure of electoral system disproportionality. John Garry is Professor of Political Behaviour in the School of History, Anthropology, Philosophy and Politics at Queen’s University Belfast. He has published widely on the themes of elections, public opinion and parties. His latest book is Consociation and Voting in Northern Ireland: Party Competition and Electoral Behaviour, published in 2016 by University of Pennsylvania Press. He was also a co-author of the prize-winning book The Irish Voter, published in 2008 by Manchester University Press. Michael Marsh, MRIA, is Emeritus Professor of Political Science at Trinity College Dublin. The author of a wide variety of articles on parties and electoral behaviour, he has co-edited each of the last five books in the  How Ireland Voted series, and has been a principal investigator for the Irish National Election Study since its foundation. He was co-author of the first book arising out of those studies, The Irish Voter (Manchester, 2008), and co-edited the next one, A Conservative Revolution? (Oxford, 2017).  Gail McElroy is Professor in Political Science at Trinity College, Dublin. Recent published work explores party competition in the Republic of Ireland, voting behaviour in preferential voting systems and political group cohesion in the European Parliament.  Eoin O’Malley is Associate Professor at the School of Law and Government, Dublin City University. His research interests broadly relate to Irish politics

Notes on contributors

xv

including cabinet government, policy-making and leadership. He is author of over forty peer-reviewed publications, and co-editor of One Party Dominance: Fianna Fáil and Irish Politics 1926–2016 (Routledge, 2017). Stephen Quinlan is Senior Researcher at the GESIS Leibniz Institute for the Social Sciences, Mannheim and Project Manager of the Comparative Study of Electoral Systems (CSES), a study that explores electoral behaviour in over forty states globally. His research focuses on comparative electoral behaviour and public opinion including turnout, elections, referendums and social media’s impact on politics. His research has been published in  Information Communication and Society, Electoral Studies, and Irish Political Studies. Theresa Reidy is Lecturer in the Department of Government and Politics at University College Cork. Her research interests lie in the areas of political institutions and electoral behaviour and her recent work has been published in Electoral Studies, Parliamentary Affairs and Politics. She is leading a European Commission-funded project on voter facilitation and engagement practices and is a co-convenor of the PSAI specialist group Voters, Parties and Elections. Jane Suiter is Associate Professor in the School of Communications at Dublin City University. She is an expert on communication, deliberation and participation and is co-principal investigator on the Irish Citizen Assembly, having also worked on other deliberative events such as the Constitutional Convention and We the Citizens. She is communications director of a COST project on populist political communication. She is also co-convener of the PSAI specialist group Voters, Parties and Elections, with an interest in direct democracy and referendums.  Liam Weeks is Lecturer in Politics in the Department of Government, and Politics University College Cork, and an Honorary Senior Research Fellow at the Department of Modern History, Politics and International Relations, Macquarie University, Sydney. His research interests include elections and electoral systems. He is author of Independents in Irish Party Democracy (Manchester, 2017), coeditor of  Radical or Redundant? Minor Parties in Irish Political Life  (Dublin, 2012) and  co-author of  All Politics is Local: A Guide to Local Elections in Ireland  (Cork, 2009). His next book (co-edited), The Treaty. Establishing Ireland’s Independence will be published by Irish Academic Press in 2018.

Editors’ preface

The Irish National Election Study (INES) of 2016 is the fourth such study of an Irish election since 2002. As we set out in the following pages, it is hard to imagine a more significant election to study – Ireland’s first post-crisis election. This book is very much a joint endeavour. Under the helpful auspices of the Political Studies Association of Ireland (PSAI), we organized a workshop in parallel with the 2016 conference of the association held in Belfast. The workshop was funded by the Irish Research Council New Foundations scheme. We are very grateful to our chapter authors for having produced draft chapters on time and to order, for their speedy responses to our requests for redrafts, and for redrafts of redrafts in the months that followed. We provide full details of the INES surveys in chapter 1 and the appendix to this volume, but here we wish to acknowledge the financial help that enabled them to happen. Unfortunately – as in 2011 – we were unable to secure research council funding to conduct the 2016 election study. But we were determined that an election study should proceed and therefore sought assistance from a wide range of sources to enable us to co-sponsor with RTÉ (the national broadcaster) the Behaviour and Attitudes/RTÉ exit poll, and to commission two separate (but related) telephone polls, carried out by RED C soon after the election. We are grateful to the following for making all of this possible: the School of Politics and International Relations at University College Dublin; the Department of Political Science at Trinity College Dublin; the Department of Government at University College Cork; the School of Politics, International Studies and Philosophy at Queen’s University Belfast; the School of Law and Government at Dublin City University; the Oireachtas (Irish parliament); and the Department of Justice and Equality.  Without these generous sources of funding, there would not have been an INES 2016. But, frankly, this model of putting out the begging bowl far and wide to support academic research of Irish elections is not sustainable. It is our fervent

Editors’ preface

xvii

hope that by the time of the next Irish elections, the academic community will have secured sufficient resources to enable a properly-funded election study to be implemented once again, and that – as in most other democracies today – this becomes the norm in Ireland. Finally, we are grateful to Tony Mason of Manchester University Press for his support and advice (and patience) at every stage of this process.

1 Ireland’s post-crisis election Michael Marsh, David M. Farrell and Theresa Reidy

The 2016 general election in the Republic of Ireland was dramatic. It delivered the worst electoral outcome for the established parties in the history of the state, the most fractionalized party system in the history of the state, the greatest number of Independent (non-party) TDs (MPs) elected to parliament in the history of the state, several new political parties and groups, and was one of the most volatile elections with among the lowest of election turnouts in the state’s history. These outcomes follow a pattern seen across a number of Western Europe’s established democracies in which the ‘deep crisis’ of the Great Recession has wreaked havoc on party systems (e.g. Hernández and Kriesi, 2015). The objective of this book is to assess this most extraordinary of Irish elections both in its Irish and wider cross-national context. With contributions from leading scholars on Irish elections and parties, and using a unique dataset – the Irish National Election Study (INES) 2016 – this volume explores voting patterns at Ireland’s first post-crisis election and considers the implications for the electoral landscape and politics in Ireland. This chapter sets the scene for the chapters that follow. We start by presenting a short background to the 2016 election, before describing the features of the 2016 INES. We then outline the key themes addressed in the book. This is followed by an overview of each of the chapters.

Background to the 2016 election The general election was held on 26 February 2016. There were a number of legislative and political developments during the 2011–2016 Dáil (Irish parliament) term which shaped the dynamics of competition at the election. The number of seats to be filled in the thirty-second Dáil was reduced to 158 following the passage of the Electoral (Amendment) (Dáil Constituencies) Act, 2013. A commitment to reduce the number of TDs had been included

2

The post-crisis Irish voter

in the programme for government as part of the political reform plans of the government elected in 2011. The reduction in numbers contributed to an intensification of competition for the available seats and was generally viewed to have harmed the government more than other groups (Gallagher and Marsh, 2016). Legislative gender quotas were introduced through the Electoral (Amendment) (Political Funding) Act, 2012. The new legislation made a large proportion of state funding for political parties contingent on their running a minimum of 30 per cent of candidates of both sexes. In practice, of course, this worked as a quota for female candidates, and Fianna Fáil and Fine Gael experienced some internal tensions in their efforts to meet this obligation. Finally, new party formation was a feature of the 2011–2016 Dáil with Renua emerging as a splinter group from Fine Gael, the Social Democrats forming out of a new alignment of Independent TDs and, similarly, Independents 4 Change emerging from an alignment of former party TDs and Independent TDs, although the latter would continue to organize as a loose coalition and did not adopt the features of a political party (for more on all this see Gallagher and Marsh, 2016; Marsh, Farrell and McElroy, 2017). Independents have long been a feature of the political system and in 2016 some of this group also adapted their operating principles, with several sitting TDs and senators 40

Fianna Fail

Sinn Fein

Fine Gael

Green Party

Labour

Independent/other

30

20

10

0 Jan-16

Jan-16 Feb-16 Feb-16 Feb-16 Feb-16 Feb-16 Feb-16

GE Feb 16

Figure 1.1  Party support levels during January and February 2016 and election result. Source: RED C Marketing and Research (http://www.redcresearch.ie/). Note: Dates used in Figure 1.1 do not relate to poll dates – they are time points for illustration.

3

Ireland’s post-crisis election

forming loose alliances, most notably the Independent Alliance. In sum, the election, when it was announced, was for a smaller Dáil, with more women candidates contesting than ever before and with several new party entrants to electoral competition. Despite speculation during the late summer of 2015 that an early election was imminent, the thirty-first Dáil saw out its full term. The Dáil was dissolved on 3 February 2016 by the Taoiseach (prime minister) Enda Kenny and the short campaign which followed was generally deemed dull and uninteresting by the attending media. Protestations of boredom aside, the campaign did matter and as can be seen from Figure 1.1, there were notable changes in the support levels for the main parties and groups during that time. Fianna Fáil and Independents trended up while Fine Gael and Sinn Féin trended down. The decline in support for Fine Gael ran contrary to a narrative in advance of the campaign that once voters engaged with important economic issues, the fortunes of the government parties would revive. The polls captured the fragmented landscape quite well and it was apparent from early in the campaign that the thirty-second Dáil would be more diverse and fragmented than ever before. The final outcome of the election is recorded in Table 1.1. Turnout was 64.5 per cent. There are a number of striking features which emerge, two of which in particular are worth noting. The combined vote share of the longstanding largest parties, Fianna Fáil and Fine Gael, dropped below 50 per cent for the first time. There was a sharp rise in the vote share for Independents and small parties, resulting in a fragmented electoral landscape. How and why voters delivered these outcomes is addressed throughout the chapters in this book.

Table 1.1  Results of the 2016 general election Party Fine Gael Fianna Fáil Sinn Féin Labour AAA-PBP Social Democrats Greens Other

% Vote Share

Dáil Seats

25.5 24.3 13.8 6.6 3.9 3.0 2.7 20.2

50 44 23 7 6 3 2 23

Source: Department of Housing, Planning and Local Government. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

4

The post-crisis Irish voter

The 2016 INES All of the analyses in this volume make extensive use of the 2016 INES. This is the fourth such study (the output of the earlier studies is presented in Marsh et al., 2008; Marsh, Farrell and McElroy, 2017). Budget limitations required a somewhat different approach to the earlier studies; accordingly, on this occasion, the INES consisted of three discrete surveys (full details of these are provided in the appendix to this volume).1 First, there was a nationwide exit poll of 4,283 voters as they left the polling station (done in conjunction with the national broadcaster, RTÉ, it was implemented by the Behaviour & Attitudes market research company). Because the exit poll was face to face, it facilitated the inclusion of a mock ballot, a crucial element in the study of voting behaviour in this most ‘candidate-centred’ of electoral systems: the single transferable vote (Farrell and McAllister, 2006). Respondents were provided with a facsimile of the ballot they had just completed moments earlier and asked to mark it as they had done. This provides us with information on the voter’s ranking of both candidates and parties. It was supplemented by questions asking voters to rate the parties, such as the standard items tapping affective orientations to parties, and questions asking respondents to rate the candidates. This provides a dataset, as with previous waves of INES, that is unusually (perhaps uniquely so) rich in terms of tapping respondents’ rating and ranking of the elements of electoral choice. By its nature, an exit poll is time delimited to no more than about 10 minutes, which restricts how many questions can be asked. To maximize the number of questions we could ask, the respondents – all of whom had first completed the mock ballot and answered a series of core questions – were divided into three subsamples, each of which were given a different set of questions. In order to broaden our analysis of voting behaviour in this election, we also commissioned two separate telephone polls of representative samples of Irish voters. These surveys were implemented by RED C. One of the polls applied a battery of questions from the latest wave of the influential Comparative Study of Electoral Systems (CSES) project (www.cses.org), which in this election round was focused on the theme of populism – a prominent theme in much of the analysis in this book. In the next section, we set out the key themes of this election (some of which are particular to this election, others the culmination of tendencies that started earlier) that guide the analysis to follow in the remaining chapters.

Key themes The objective of the book is to provide a comprehensive understanding of voter decision making in Ireland while also assessing some of the theoretical

Ireland’s post-crisis election

5

propositions that currently dominate debates on elections in established democracies. From an Irish perspective, six propositions drawn from the international literature are particularly pertinent: changing partisan identities, issue mobilization, new ideological dimensions of politics, party system change, populism and generational effects. Structuring the analysis along these lines facilitates both a deep analysis of the 2016 election and the location of the Irish experience in a cross-national perspective. Changing partisan identities: Partisanship has long been low in Ireland by international standards, yet elections have tended to produce relatively stable outcomes over many decades. The year 2016 was an especially volatile election and the final result raises interesting questions about party attachment. As we have noted, several new parties emerged between 2011 and 2016; their capacity to establish enduring partisan ties will be important for their longterm prospects. The reversal of Fianna Fáil’s decline and the consolidation of Sinn Féin support hint at changing partisan identities, certainly in the latter case. The relationship between voters and parties is addressed across chapters 2, 5, 7, 9 and 10. Issue mobilization: A number of controversial policies were introduced during the 2011–2016 parliamentary term, most notably water charges, which prompted huge protests across the state and a widespread refusal to pay (leading to a decision by the government elected in 2016 to abolish the charges). The high-profile marriage equality referendum in 2015 led to a surge in voter registration and turnout jumped sharply from previous referendums (Elkink et al., 2017). Taken in the round, it is clear that issues have the capacity to mobilize voters, and how these types of issues manifested at the 2016 election will be considered across chapters 3, 5, 7, 8 and 10. Ideological dimensions: Voters in many established democracies are increasingly distrustful of, and indeed angry at, politicians and political parties. This situation is sometimes attributed to economic uncertainty and vulnerability caused by hyper-globalization, the idea being that globalization has created a different set of economic winners and losers and so has activated a new and enduring political cleavage. A social dimension has also been mooted, one that suggests voter attitudes can increasingly be interpreted along a localistinternationalist spectrum. Ireland, like so many small open economies, does not fit easily into these lines of analysis. Despite the deep economic crisis from 2008, globalization has brought many improvements, most especially increased employment. But not all parts of the country have benefitted equally – Dublin, the engine of the national economy, has powered ahead, leaving a strong perception of regional imbalance in the economic recovery. When we intersect the long-standing localist disposition of Irish voters with the complexity of twenty-first-century economics, we may find that the tectonic plates of politics are shifting. Left and right, never especially powerful in underpinning voter

6

The post-crisis Irish voter

preferences, and conservative and liberal, may be evolving into new dimensions; the question of the precise substance and shape of these dimensions and how likely they are to endure will be considered across chapters 3, 4, 5, 6 and 10. Party system change – hollowing of the centre: The party-political landscape in Ireland has certainly changed. The vote share of the traditional parties shrank below 50 per cent for the first time in 2016 and new parties gained ground. Candidate gender quotas were introduced in 2012 and required all parties to offer voters a greater choice. Chapters 5, 7, 9 and 11 will develop this analysis across the book and emphasize the implications of 2016 voter behaviour for the party system. Populism: Many countries have seen a surge in anti-establishment sentiment and consequent growth in support for populist political parties. Distrust of politicians and political parties is high in Ireland, but this feature of anti-politics is moderated by localist considerations and candidate-centred aspects of competition. The introduction of water charges begot a large protest movement which was subsequently integrated into, in part, a coalition of parties on the left. Chapters 3, 7 and 10 will investigate this development, paying particular attention to what populism means in an Irish context and proceeding to evaluate the support base of new parties contesting elections for the first time in 2016 and of Independents. Generational effects: There is plenty of international evidence of the impact of generational change on voting behaviour. It is unlikely that Ireland will be immune from this. We can anticipate generational differences in a number of respects, including differing values and attitudes towards politics, perhaps a greater focus on national policy questions (over localist, particularist concerns) and variations in degree and forms of political involvement (e.g. greater use of social media). In addition, the differential impact of the recession across the generations could also feed into variations in voting behaviour. This issue features most in chapters 2, 5, 8 and 10. The ten chapters that follow track the behaviour of the post-crisis Irish voter in this 2016 election in the light of these themes. In the next section, we briefly review the focus of each of the chapters.

Overview of chapters The single transferable vote (STV) electoral system used for elections in Ireland provides a valuable data resource for political scientists. Because it invites voters to rank order as many (or as few) candidates as they like on the ballot paper, it provides important information on voting patterns beyond the first preference expressed on the ballot paper. In chapter 2, Kevin Cunningham makes use of mock-ballot data gathered as part of the exit poll (INES1) of voters as they were leaving the polling station. This allows him to examine the stability in first

Ireland’s post-crisis election

7

preference voting behaviour in 2016 and how this has changed since before the financial crisis. The chapter also explores the patterns of lower preferences and what they might mean for the party system. Finally, Cunningham addresses whether preferences mattered in terms of the number of seats a party won in 2016. Cunningham’s main findings are twofold. First, he notes that the erosion of party allegiances that was so evident in the 2011 election (Marsh, Farrell and McElroy, 2017) has continued. Even though the worst of the financial crisis had abated, large numbers of voters continued to switch votes from one party to another in 2016. Second, there is the intriguing finding of the emergence of two parallel party systems in terms of the transfer of voter preferences, with voters on the right transferring votes between the main established parties while those on the left transferred between non-established parties (principally Sinn Féin and AAA-PBP). This latter trend has important implications for Sinn Féin, which in the past tended to be ‘transfer toxic’, but perhaps is no longer so, at least to supporters of parties of the left. The evidence of change in party allegiances raises important questions about the ideological positioning of parties in 2016, which is the focus of chapter 3. One of the classic features of party politics in Ireland – that helps to mark it out as ‘unique’ (Whyte, 1974; Carty, 1981) – is the lack of ideological distinction between the parties, particularly relating to the two main established parties, Fianna Fáil and Fine Gael. In this chapter, Kevin Cunningham and Johan A Elkink evaluate the extent to which ideology may now matter more than before. They do so by analysing the relationship between the ideological positions of parties and vote choice, and by developing a dimensional mapping of ideological space based on rankings in the mock-ballots. Their principal conclusion is that while it may still be the case that ideology does not play a lead role in Irish politics, perhaps now, in their words, it might be seen at least as ‘a supporting actor’. It remains the case that ideological positioning does not separate the two largest Irish parties, Fine Gael and Fianna Fáil; however, ideology does determine whether someone might vote for either or neither of these parties. On average, Irish voters select parties that are ideologically close to them on a left–right scale, most prominently so for voters on the left of the spectrum where left versus right does matter in their choice between parties. Cunningham and Elkink’s analysis reveals a significant overlap between socio-moral, economic left and right and globalization issues, but there is evidence that populism is offering a new cross-cutting dimension defining political competition. Overall, from a comparative perspective, the Irish case may appear more conventional in terms of left–right competition than typically assumed; it also has an undercurrent of anti-globalization that is similar to that found in other European states.  In chapter 4, John Garry delves more deeply into the social and ideological bases of voting behaviour in 2016. Referring to the classic debates in the

8

The post-crisis Irish voter

comparative literature on political cleavages (Lipset and Rokkan, 1967) and earlier empirical investigations of the Irish case (notably by Marsh et al., 2008), the core question he seeks to answer is whether there may be a strong link between voters’ socio-demographic traits, their broad policy beliefs and their party choice in this election. Building upon a similar study of the 2011 election, which found evidence of the emergence of class-based politics (Tilley and Garry, 2017), the analysis on this occasion reveals some interesting trends, particularly relating to Sinn Féin. Its steady rise in electoral support over time has seen it emerge as a major player in Irish party politics, with important implications for how we might view the ideological basis of voting behaviour in Ireland. Garry’s analysis finds that Sinn Féin’s strong socio-demographic profile (working class, left-wing and in favour of Irish unity) sets it apart from the other major parties, differentiating it in terms that would be familiar in a political cleavage-based analysis. When voters form attachments to political parties, this can colour their perceptions of candidates, policies and events. As Campbell et al. (1960) put it, ‘Identification with a party raises a perceptual screen through which the individual tends to see what is favourable to his partisan orientation.’ Party attachments therefore tend to be immune to all but the most seismic political shocks (Green et al., 2004). The economic crisis that engulfed Ireland in 2008 may represent such a shock, leading as it did to a period of unprecedented electoral volatility. In chapter 5, Rory Costello presents the first dedicated study of party attachment in Ireland in the wake of the economic crisis. Previous research shows that party identification has historically been an important factor in Irish voting behaviour (Carty, 1981). Partisanship began to decline from the 1980s onwards, but Marsh et al. (2008) were still able to report that the majority of voters had some degree of party attachment at the beginning of the millennium. This chapter examines how party attachment has evolved in recent elections. The core question Costello seeks to answer is whether the electoral turbulence in 2011 and 2016 was simply a symptom of a fundamentally dealigned electorate, or whether we are witnessing a realignment in Irish politics. In other words, has the number of floating voters increased in the wake of the crisis, or have people begun to form new party attachments that are likely to shape elections in the future? The analysis shows that while party attachments were ruptured in 2011 (most notably so in the case of Fianna Fáil), in 2016, by contrast, partisanship increased, and there were some interesting trends among young voters in particular, with many of them beginning to form new allegiances. In chapter 6 Michael Marsh seeks to explain a significant puzzle of the 2016 election. There is now a very extensive literature linking economic performance with the electoral performance of government parties, with the relationship being a positive one. The 2016 election was an unusual illustration of a government being punished despite being able to point to a record of very

Ireland’s post-crisis election

9

significant economic growth and rapidly falling unemployment as Ireland’s recovery from the economic crash and bailout made it such a good example of the success of ‘austerity’ policies. Drawing on many studies that argue for certain contingencies in the relationship, this chapter explores a number of ways in which the ‘good economy-government returned to office’ relationship went wrong. A key finding, contrary to general tendencies in the literature on economic voting, is that ‘pocketbook’ considerations were very significant in determining how voters felt about the government parties. The chapter offers some reasons why the Irish case is unusual and also questions the theoretical bases on which ‘pocketbook’ voting is downplayed in the economic voting literature. In chapter 7, the focus turns to the significance of candidates in Irish elections. The act of voting is often judged to be party-centred (Elkink and Farrell, 2016), but in Ireland it is generally seen as taking place through the prism of candidates: parties select their candidates with care to take account of that; candidate-centred behaviour is also shown by the large and growing number of Independents elected in recent Irish elections – in record numbers in 2016. The importance of party versus candidate has been examined in previous studies (Marsh, 2007; Marsh et al., 2008). In this chapter, Michael Courtney and Liam Weeks bring the discussion up to date for 2016. The financial crisis had a number of political impacts, and one was to increase the importance of party vis-à-vis candidate in 2011 (Marsh, Farrell and McElroy, 2017). This was because national issues, which parties are more capable of dealing with than individual candidates, became of greater importance. With the gradual recovery of the Irish economy in the latter half of the tenure of the Fine Gael–Labour coalition, Courtney and Weeks consider whether the dynamics of party and candidate were altered. Their analysis shows that voters have returned to the more familiar habit of candidate-centred ballot choices, though significant party-centred behaviour persists. Chapter 8 concentrates on when voters make their voting decisions, paying particular attention to the campaign period. Theresa Reidy and Jane Suiter argue that knowing when decisions are made is a vital part of understanding how elections work. The evidence presented demonstrates that a growing proportion of voters report making their final voting choice during election campaigns. Modern election campaigns, with their manifesto launches, party leader debates and intense scrutiny of opinion polls, matter a great deal. These campaigns work by raising awareness of new parties and candidates and providing vital information on the policy positions of competing actors. Reidy and Suiter’s analysis reveals that they are decisive in shaping voter decisions. Young people, women and urban voters are more likely to arrive at their final vote choice during the campaign period and, importantly, voters who decide during the campaign are also more likely to have changed their preference

10

The post-crisis Irish voter

from the previous election. These findings have important implications for the political system. They provide further evidence of the challenges parties face in building long-term allegiance with their voters. Furthermore, it is clear that election results may become more unpredictable as larger proportions of voters arrive at their final decision close to election day, making early campaign opinion polls more problematic as predictors of final outcomes. In 2016, Ireland joined over fifty countries worldwide in the adoption of candidate gender quotas, and it became the first case of a country doing so under the STV electoral system. Its impact was evident from the dramatic rise in the number of women candidates fielded in this election – 163, as compared with 86 in 2011. In chapter 9, Gail McElroy builds on her previous research of the Irish case (McElroy and Marsh, 2010; 2011) to assess whether the use of gender quotas had any impact on voters’ attitudes towards women candidates. Her analyses of INES data in previous elections found no evidence of voter prejudice against female candidates. There could be reason to expect that this might change in the light of gender quotas. As McElroy observes, the introduction of the quota in 2016 was a significant ‘shock’ to the system: parties were forced to find a large number of women candidates very quickly, so the recruitment pool was likely to have more ‘average’ women in it. This allows her to test for true bias among the Irish electorate. Her analysis reveals little evidence of this on the whole, apart from the slight exception of Fianna Fáil, whose supporters revealed some male bias. Apart from that partial exception, the findings generally are consistent with previous studies: what matters most is how well the candidate is known, and therefore it is incumbency that is the main factor, not the sex of the candidate. In chapter 10, David Farrell, Michael Gallagher and David Barrett assess how the record-breaking levels of electoral flux in 2016 may have impacted on attitudes towards representative politics in Ireland. The chapter addresses three themes. First, they examine voter attitudes to the role of TDs in 2016. The Irish tradition of high degrees of localism in representative politics is based on the strong attachment of Irish voters to a constituency orientation from their politicians (Gallagher and Suiter, 2017; Gallagher and Komito, 2018). While there may be grounds for expecting that the high levels of electoral change in recent elections may have had some impact on this, the analysis in this chapter shows that a desire for a constituency orientation remains as strong as ever. There are, however, some changes in how voters make contact with their elected representatives – the second theme dealt with in this chapter. The intensity (or degree) of contact is resilient, but its form is shifting to more impersonal or virtual means of contact (especially among younger voters): the days of the ‘weekly clinic’ – that classic mainstay of representative politics in Ireland – may be numbered. Finally, the chapter examines what Irish voters think of their politicians overall – this latter theme referencing ongoing international debates

Ireland’s post-crisis election

11

about the emergence of populist attitudes (which were central to this round of CSES questions, asked in the INES 2016). The evidence from the Irish case is quite positive, with many voters indicating a favourable disposition towards their politicians. This is not universal, however; supporters of Sinn Féin and AAA-PBP are somewhat more critical of politicians. The final chapter (chapter 11) addresses the importance of leadership effects in 2016. Stephen Quinlan and Eoin O’Malley assess the impact of the leaders of the four main parties in influencing the vote for their parties. Overall, they find some evidence that party leadership mattered in this election, but not a lot. The Fianna Fáil leader, Micheál Martin, was the most popular of the leaders yet this did not translate into significant additional votes for his party. By contrast, the leaders of Fine Gael (Enda Kenny) and Sinn Féin (Gerry Adams), though less popular, were better at influencing the turnout of their base of supporters.

Conclusion Late in the evening of the election count on 27 February 2016, as it became apparent that the result was one in which no party could claim a success (Gallagher and Marsh, 2016), one of the hosts on the RTÉ election results programme speculated whether Ireland was about to enter a period of Borgenstyle government – referencing the popular Danish political drama about a prime minister who, against all the party-political odds, managed to cobble together a minority coalition government. It was a prescient observation. Government formation was to take several months and the final deal, when agreed, returned Fine Gael to power in a minority administration with the involvement of a number of Independent TDs, including some Independents from the Independent Alliance group. The government is reliant on a confidence-and-supply agreement with Fianna Fáil which commits that party to abstaining on key legislation, and the deal covers three budget years. The speculation from the beginning of this government has been that it will not be long-lasting, with many expecting that the next election will be no later than 2019. But so far, the government has endured. At the time of writing (late autumn 2017), the government has successfully passed its second budget and the Fine Gael party has changed leader, resulting in the election by the Dáil of Leo Varadkar as Taoiseach. All of this has the familiar feel of political life as normal. In short, while there was much that was extraordinary about this election and its outcome, there is also much that remains the same about Irish politics.

Note 1 The INES 2016 received funding from the following sources: the School of Politics and International Relations at University College Dublin; the Department of Political Science

12

The post-crisis Irish voter

at Trinity College Dublin; the Department of Government at University College Cork; the School of Politics, International Studies and Philosophy at Queen’s University Belfast; the School of Law and Government at Dublin City University; the Oireachtas (Irish parliament); and the Department of Justice and Equality.

References Campbell, Angus, Philip E. Converse, Warren E. Miller and Donald E. Stokes. 1960. The American Voter. New York: John Wiley and Son. Carty, R. Kenneth. 1981. Electoral Politics in Ireland: Party and Parish Pump. Cork: Brandon. Elkink, Johan, and David Farrell (eds). 2016. The Act of Voting: Identities, Institutions, and Locale. London: Routledge. Elkink, Johan, David Farrell, Theresa Reidy and Jane Suiter. 2017. ‘Understanding the 2015 marriage equality referendum in Ireland: Context, campaign and conservative Ireland’, Irish Political Studies. 32: 361–81. Farrell, David, and Ian McAllister. 2006. Australia’s Electoral Systems: Origins, Variations and Consequences. Sydney: University of New South Wales Press. Gallagher, Michael, and Jane Suiter. 2017. ‘Pathological parochialism or a valuable service? Attitudes to the constituency role of Irish parliamentarians’, in Michael Marsh, David Farrell and Gail McElroy (eds), A Conservative Revolution? Electoral Change in Twenty-First Century Ireland. Oxford: Oxford University Press, pp. 143–71. Gallagher, Michael, and Lee Komito. 2018. ‘The constituency role of Dáil deputies’, in John Coakley and Michael Gallagher (eds), Politics in the Republic of Ireland. 6th edition, London: Routledge, pp. 191–215. Gallagher, Michael, and Michael Marsh (eds). 2016. How Ireland Voted, 2016: The Election that Nobody Won. London: Palgrave Macmillan. Green, Donald P., Bradley Palmquist and Eric Schickler. 2004. Partisan Hearts and Minds: Political Parties and the Social Identities of Voters. New Haven, CT: Yale University Press. Hernández, Enrique, and Hanspeter Kriesi. 2015. ‘The electoral consequences of the financial and economic crisis in Europe’, European Journal of Political Research (published online, 2015; accessed 28 April 2016). Lipset, Seymour M., and Stein Rokkan. 1967. Party Systems and Voter Alignments: Cross-National Perspectives. New York: Free Press. Marsh, Michael. 2007. ‘Candidates or parties? Objects of electoral choice in Ireland’, Party Politics. 13: 500–27. Marsh, Michael, David M. Farrell and Gail McElroy (eds), 2017. A Conservative Revolution? Electoral Change in Twenty-First-Century Ireland. Oxford: Oxford University Press. Marsh, Michael, Richard Sinnott, John Garry and Fiachra Kennedy. 2008. The Irish Voter: The Nature of Electoral Competition in the Republic of Ireland. Manchester: Manchester University Press.

Ireland’s post-crisis election

13

McElroy, Gail, and Michael Marsh. 2010. ‘Candidate gender and voter choice: Analysis from a multimember preferential voting system’, Political Research Quarterly. 63: 822–33. McElroy, Gail, and Michael Marsh. 2011. ‘Electing women to the Dáil: Gender cues and the Irish voter’, Irish Political Studies. 26: 521–34. Tilley, John, and John Garry. 2017. ‘Class politics in Ireland: How economic catastrophe realigned Irish politics along economic divisions’, in Michael Marsh, David Farrell and Gail McElroy (eds), A Conservative Revolution? Electoral Change in Twenty-First Century Ireland. Oxford: Oxford University Press, pp. 11–27. Whyte, John, 1974. ‘Ireland: Politics without social bases’, in Richard Rose (ed.), Electoral Behaviour: A Comparative Handbook. New York: The Free Press, pp. 619–51.

2 Mining the ballot: Preferences and transfers in the 2016 election Kevin Cunningham

Introduction The purpose of this book is to explain Irish voting behaviour (why people in Ireland vote the way that they do) after the crisis. The 2016 general election was the first since the ‘earthquake’ election of 2011, which saw a historic change in the Irish party system. While the 2009 and 2014 European elections pointed to a shift in voting patterns, the second-order effects apparent in those elections (see Marsh, 1998; Hix and Marsh, 2011) would blur our understanding of developments since 2008. Thus, the 2016 general election is the first test of permanency of a shift in voting behaviour since the economic crisis that began in 2008. To understand the post-crisis voter, this chapter analyses first preference votes cast for each political party, and the stability of those preferences over time and across parties. The chapter explores the additional data produced by Ireland’s distinctive electoral system – the single transferable vote (STV). Voters are invited to rank candidates on the ballot paper, marking a ‘1’ beside the candidate they most prefer, a ‘2’ beside their second most preferred candidate, a ‘3’ beside their third most preferred candidate and so on. Voters can rank any number of candidates, from just one candidate to all candidates. The count proceeds in rounds eliminating the candidates with the fewest votes and redistributing their votes according to lower preferences. Candidates are certain to be elected if they achieve a quota of votes as determined by the number of votes cast and the number of candidates to be elected (although when the count is concluded, some will be declared elected simply because they have more support than any other remaining candidate). Lower preferences are thus an important part of a voter’s considerations and therefore information about these lower preferences gives us a greater insight into the thought processes underlying voting behaviour.

Mining the ballot

15

A number of different rubrics may drive the pattern of preferences on the ballot. It may be that the voter considers the political parties they like while avoiding political parties they explicitly do not like. They may favour governing parties or opposition parties, parties of the left or, equivalently, of the right. Voters may favour specific candidates due to local or personal factors. Other thought processes such as convenience might dictate a significant number of preferences, with voters favouring candidates higher up on the ballot paper. This chapter discusses the post-crisis Irish voter, first by looking at the instability in first preference voting behaviour and how this has changed since before the crisis. It also explores trends among the lower preferences – how many parties voters are voting for, and the patterns of lower preferences – and what it might mean for the party system. Finally, we explore whether preferences mattered in terms of the number of seats the party won.

Stability of preferences Until 2011, the Irish political system was known for its somewhat unusual stability. While support for the individual parties changed quite significantly from one election to the next, the party system did not change very much. Between 1927 and 2007 the average absolute change in the Fianna Fáil and Fine Gael vote from one election to the next was 4 per cent. However, as Figure 2.1 shows, support levels tended to oscillate in a range that left intact the basic structure of the party system: in eighteen post-war elections, the order and relative sizes of the three largest parties remained the same. Fianna Fáil was consistently the largest party, typically receiving between 39 and 52 per cent of the vote; Fine Gael was the second largest with support ranging between 19 and 39 per cent; and the Labour Party was third (in all but one election), with support ranging between 6 and 19 per cent of the vote. As Figure 2.1 also shows, the 2007 election was the last election to be characterized by this pattern of stability. Since the economic crisis, the party system has entered a period of instability or at least realignment. The average change in vote share increased from 3.7 per cent to 14.1 per cent in 2011 and 10.1 per cent in 2016. The 2011 election is ranked as the third most volatile election in long-established European democracies since 1945 and as the most volatile without the presence of a significant new party (Mair, 2011: 287–88). While the 2016 election saw less change than 2011, it was still the second most volatile of all Irish elections. The 2011 election was characterized by a dramatic collapse in support for the government, and 2016 was no different. Despite considerable economic gains over the previous five years, support for the two governing parties fell by 23 per cent in total. The 2016 election was thus ranked as the eighth most volatile of all elections in Western Europe since 1945 (Farrell and Suiter, 2016: 279–83).

16

The post-crisis Irish voter

Figure 2.1  First preference vote shares since 1922. *Note that Figures Fine Gael relate to Pro-Treaty Sinn Féin in 1922 and Cumann na nGaedheal from 1923 to 1933. Fianna Fáil relates to Anti-Treaty Sinn Féin in 1922 and Republican Sinn Féin in 1923. Source: ‘ElectionsIreland’, http://www.electionsireland.org 1998–2017, Christopher Took and Seán Donnelly.

To understand this post-crisis voter, the analysis begins by looking at trends in individual voting behaviour between 2011 and 2016. Firstly, respondents in INES11 were asked whether they voted in the previous election and, if they did vote, which party they supported. Using this information and what we know about the overall vote shares of the 2011 and 2016 general elections we can estimate the matrix in Table 2.1 of where voters came from and who they voted for in the 2016 election. From Table 2.1 we can summarize the number of people switching, the number of people staying and proportions of non-voters across the two elections. Approximately 29 per cent switched from one party (including Independents as one group) to another (including Independents) while almost the same proportion, 28 per cent, stayed with their party of choice (or remained with Independents) between the two elections. This is a dramatic change since before the crisis. We contrast this with data from the 1997–2002 electoral cycle (studied by Marsh et al., 2008), when a larger 34 per cent remained with their party of choice, while just 11 per cent switched from one party to another. This represents a decline from three out of four of those who did vote staying with their party of choice down to one in two. While volatility in the post-crisis era consisted of people switching between parties, in the pre-crisis era many more voters appear to make more deliberate decisions as to whether they would vote or not. Between 2011 and 2016, 23 per cent stayed at home on both occasions, whereas 20 per cent voted in only one election. In the period 1997–2002, while a similar proportion (22 per cent)

Table 2.1  Voting and non-voting in 2016, by voting and non-voting in 2011  

2011 Vote  

Fine Gael Fine Gael

2016 Vote

Fianna Fáil

Fianna Fáil

Sinn Féin

Labour

Green

Independent

Other

Did not vote

Total

11.5

0.5

0.2

1.7

0.1

0.6

0.2

1.8

16.6

3.4

7.2

0.3

1.4

0.1

1.0

0.7

1.6

15.8

Sinn Féin

1.0

0.5

3.5

1.6

0.1

0.5

0.2

1.6

9.0

Labour

0.9

0.2

0.0

2.2

0.0

0.4

0.1

0.5

4.3

AAA-PBP

0.4

0.0

0.0

0.8

0.2

0.0

0.6

0.7

2.6

Social Democrats

0.2

0.1

0.0

1.0

0.1

0.3

0.0

0.2

2.0

Green

0.2

0.0

0.0

0.7

0.3

0.1

0.0

0.5

1.8

Independent

2.1

1.2

0.6

2.5

0.0

3.6

0.2

1.0

11.3

Other

0.6

0.2

0.2

0.5

0.1

0.3

0.0

0.1

1.9

Did not vote

4.5

2.1

2.2

1.0

0.5

1.4

0.0

23.3

34.9

25.2

12.2

6.9

13.6

1.3

8.5

2.1

30.1

100.0

Total

Note: Author’s analysis of INES3 and 2016 election results, N=1000. Data in the table are iteratively adjusted to ensure that the margins correctly reflect the results in 2011 and 2016. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

18

The post-crisis Irish voter

stayed at home on both occasions, a much larger proportion (33 per cent) voted in only one election. This analysis is reliant on two electoral cycles and the ability of a respondent to recall who they voted for. It is also subject to recall bias, where respondents may be more likely to recall past preferences in line with current preferences. This does mean that the proportion of respondents changing from one party to another may be even greater than the 29 per cent reported here (see Cunningham and Marsh, 2018). The panel design of the Irish National Election Study covering the period 2002–2007 asked respondents about their vote in 2002 and again in 2007 and so did not have to rely on the memory of the respondent. Sixty-three per cent of the sample had the same preference (party or Independent) at the start and end of the period (author’s analysis of data from the INES 2002–2007). Given what we know about recall bias, this is more in line with the self-reported three in four voters staying with their party of choice between 1997 and 2002, rather than the self-reported one in two between 2011 and 2016. Table 2.2 breaks this down by party of choice and, in contrast with the 1997–2002 period, it shows that the rise in the number of people switching to another party, or ‘defecting’, was consistent across all parties and Independents, whether they were rising or falling in terms of their support levels between 2011 and 2016. Naturally, as a party experiencing vote loss, the Labour Party had the largest increase in defectors (up from 25 per cent between 1997 and 2002 to 77 per cent between 2011 and 2016). Of the 2011 Labour Party supporters, few stayed at home in 2016, meaning that the effect of the fall in support for the Labour Party was to be more dramatic than otherwise in the constituencies Table 2.2  The proportion of voters switching from their party of choice, staying with their party and not voting from one election cycle to the next, 2011–2016 and 1997–2002 (per cent)    

2011–2016

1997–2002

Stayed Voted for Did not Stayed Voted for Did not with party another party vote with party another party vote

Fine Gael

47

35

18

54

18

28

Fianna Fáil

59

23

18

62

12

26

Sinn Féin

51

19

31

81

13

6

Labour

16

77

7

44

25

31

Independent

44

39

16

46

33

21

Did not vote

-

22

78

-

43

57

Source: Author’s analysis of data: 1997–2002 from Marsh et al. (2008) and 2011–2016 from INES3. Note: Shading used highlights higher values.

Mining the ballot

19

in which it ran candidates. However, what is perhaps more surprising is that the proportions of defectors from Sinn Féin and Fianna Fáil were also greater in 2016 despite both parties increasing their support levels between 2011 and 2016. This is even more remarkable considering the low ebb Fianna Fáil found itself in in 2011. While decisions about party preferences have become more volatile, decisions over whether or not to vote seem to have become more consistent. Seventy-eight per cent of those who did not vote in 2011 remained non-voters in 2016. This compares with the 57 per cent who stayed at home on both occasions between 1997 and 2002. The greater consistency of non-voters in 2011–2016 is even more remarkable considering that the fall in turnout was considerably larger between the 1997 and 2002 elections (76.5 per cent to 62.6 per cent) than between the 2011 and 2016 elections (69.9 per cent to 65.1 per cent), as one would expect a greater proportion of those who did not vote in 1997 not to have voted in 2002, given that many more people did not vote in 2002.2 To understand more about the pattern of voting one can look at the pattern of first, second and lower preferences in 2016. Extensive research on Irish voting behaviour has used aggregate material at a constituency level available from official results (see Gallagher, 1978; 1992; 1999; 2003; Sinnott, 1995; Sinnott and McBride, 2011; O’Kelly, 2016; Elkink and Farrell, 2016). This analysis tells us how many votes typically transfer from one candidate to another but is limited to the situations where a candidate is eliminated or when a candidate has a surplus. It therefore obscures our understanding of what is intended by voters when they cast their vote. In this instance, our analysis draws on mock-ballot data. The RTÉ/INES exit poll (INES1), like previous rounds of the INES, asked voters to replicate the preferences indicated on their ballot and we can use these data to infer the exact nature of the preferences.

Choice options and number of preferences The number of different parties that voters express a preference for gives us an idea of the nature of any diffusion of party preferences. This is particularly relevant in the context of a significant increase in the number of political parties achieving representation. Table 2.3 shows the number of parties voters express preferences for. This is compared with the number of different parties standing in the voter’s constituency and against the voter’s first preference vote. In some cases, supporters of political parties express preferences for only one party. This custom, sometimes known as ‘plumping’, is practised by the most loyal of supporters, for whom no other party is worthwhile voting for. Fianna Fáil voters, followed by Sinn Féin and Fine Gael voters, are most likely to plump for their chosen party. Just 15 per cent gave a preference to just one party; the remainder expressed further preferences for other parties. This is

Table 2.3  Number of different parties voted for by number of options (per cent) and first preference choice  

Number of different parties voted for

 

Number of parties/groupings standing  

1 2 3 4 5 6 7 8 9 10 11 Total Avg. N

Party first preference

 

5

6

7

8

9

10

11

FF

FG

SF

Lab

GP

Ind

AA

All

24 28 27 6 14 100 2.6 78

20 48 22 4 2 3 100 2.3 94

22 36 27 10 2 1 4 100 2.5 820

18 33 31 10 4 1 1 3 100 2.7 520

11 28 36 13 5 2 1 0 4 100 3.1 1309

15 30 32 11 5 3 1 1 0 3 100 2.9 910

12 26 30 15 7 5 1 1 0 1 2 100 3.2 552

23 33 29 8 3 1 2 0 1 0 100 2.5 905

17 33 28 10 4 2 2 1 2 1 0 100 2.9 1064

18 33 32 9 5 2 0 0 0 0 0 100 2.6 686

6 27 38 16 6 1 1 1 2 0 0 100 3.2 304

3 14 35 25 10 4 2 2 2 2 1 100 3.9 153

13 32 32 12 4 2 2 1 1 0 0 100 2.9 656

9 24 35 17 6 3 0 0 2 1 0 100 3.2 201

15 31 32 12 5 2 1 1 1 1 0 100 2.9 4283

Source: INES1 2016. Note: Shading used highlights higher values. Ind = Independent; FF = Fianna Fáil; FG = Fine Gael; GP = Green Party; Lab = Labour; SF = Sinn Féin.

Mining the ballot

21

lower than previously, but only marginally so (it compares with 18 per cent in 2002). In 2002, 52 per cent indicated a preference for one or two different parties whereas now this figure is down – again marginally so – to 46 per cent. While the number of parties achieving representation in the Dáil has increased, the number of parties voters express a preference for has only marginally increased. In all constituencies, voters had a minimum choice of four parties or groupings as Fianna Fáil, Fine Gael, Sinn Féin and the Green Party stood candidates in all constituencies. Yet, just 7 per cent of those that could mark a preference for more than five parties did so. In total, the average number of parties that voters express a preference for has increased marginally from 2.8 to 2.9. The data also show the extent to which tactical considerations play an important role in a voter’s considerations. Table 2.3 reveals considerable differences in the number of preferences expressed by supporters of different parties. While 23 per cent of Fianna Fáil supporters plumped for Fianna Fáil, just 3 per cent of Green Party supporters voted only for the Green Party. These figures are not influenced by the numbers of parties on offer as both parties stood in every constituency. Looking across all parties, there is some indication that voters are strategic in terms of the number of preferences that they express. Green, Labour and Anti-Austerity Alliance-People Before Profit (AAA-PBP) supporters tend to indicate more preferences, perhaps in recognition of the relatively greater likelihood that their candidate would be eliminated and their vote might be more likely to transfer to another party. These parties also ran fewer candidates meaning that whereas a voter could express preferences for two or three Fianna Fáil or Fine Gael candidates, they would not have been able to do so for, say, the Green Party, which ran a single candidate in all constituencies. In comparison to those supporting Fianna Fáil and Sinn Féin, Fine Gael voters tended to express preferences for a larger number of other parties. This is perhaps explained by public encouragement between Fine Gael and Labour politicians to express second preferences for one another as coalition partners.

Polarization of lower preferences Lower preferences tell us not just about the parties that voters like but also about the parties that voters actively dislike. These characteristics are not necessarily the inverse of one another, as some parties polarize opinion (e.g. Sinn Féin) more than others (e.g. the Green Party). Table 2.4 shows the proportion of cases where each party is awarded a voter’s first, second, third, fourth, fifth, sixth, lower or no preference, excluding constituencies where the parties do not field a candidate. Broadly speaking, second and third preferences follow the aggregate pattern of first preferences; however, some parties such as Labour

22

The post-crisis Irish voter

Table 2.4  Distribution of successive party preferences excluding cases where the parties did not stand   First preference party Second preference party Third/fourth preference party Fifth/sixth preference party Lower preference party No preference

Fine Gael 25 12 10 2 1 50

Fianna Fáil 21 9 11 3 2 54

Sinn Féin Labour AAA-PBP 16 6 6 1 3 68

8 8 15 3 2 64

7 7 6 2 3 75

Source: INES1 2016. Note: Shading used highlights higher values. Note that in 13 per cent of cases Labour did not stand a candidate and in 29 per cent of cases the AAA-PBP did not stand a candidate. Sample size: 4,283. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

attract relatively more second, third and fourth preferences than, say, Sinn Féin, which incidentally attracts twice the number of first preference votes. One may conclude that Sinn Féin and the AAA-PBP are more disliked by voters. Both are awarded lower and no preferences (71 per cent, 78 per cent) more often than Fianna Fáil (56 per cent), Fine Gael (51 per cent) and Labour (66 per cent). As this figure is related to the overall support levels of the parties, we may exclude first preference votes to get a clearer picture of which parties are disliked most. Of those that do not express a first preference for Sinn Féin, 81 per cent do not give any preference to the party. Similarly, for the AAA-PBP, of those that do not give the party a first preference vote, 81 per cent do not give the party a lower preference. This contrasts with figures of 67 per cent, 68 per cent and 70 per cent respectively for Fianna Fáil, Fine Gael and Labour. These figures reflect the considerable change in voting behaviour between the pre-crisis voter and the post-crisis voter. Pre-crisis, most voters expressed a preference for each of the main parties and a minority left these parties entirely off the ballot. In 2002, just 28 per cent did not express a preference for Fianna Fáil. In 2016, 54 per cent left Fianna Fáil entirely off their ballot. While the party is much less popular than it was in 2002, the decline is remarkable even considering the scale of its vote collapse in 2011. This 54 per cent is not too dissimilar from the 59 per cent that did not express a preference for the Labour Party in 2002, which was a much smaller party with only 13 per cent of the vote. For Fine Gael, in 2016, 47 per cent did not issue a preference for the party. Despite increasing its vote share since 2002, fewer people in 2016 expressed a preference for the party. In 2002 the main parties were more popular and had a much greater share of first preference votes. Clearly, parties with lower first preference votes will attract fewer lower preferences.

Mining the ballot

23

When we exclude first preferences for each party, the difference between the pre-crisis and post-crisis voter is still observable. Among those that did not express a first preference vote for Fianna Fáil, just 47 per cent expressed no lower preference in 2002 compared with 67 per cent in 2016. For Fine Gael in 2002, 60 per cent of those that did not vote for the party first did not express any preference for the party. In 2016 this rose to 68 per cent. Among non-Labour voters, 66 per cent left the party entirely off the ballot in 2002, a figure that rose to 70 per cent in 2016. With respect to Sinn Féin, however, the electorate has softened marginally. In 2002, 85 per cent of those whose first preference vote was not Sinn Féin left the party entirely off the ballot. In 2016 this fell to 68 per cent. While voters selected marginally more parties on the ballot, the main parties were left entirely off the ballot considerably more often. A common theme is that mainstream parties have lost their total dominance of the party system. This is something we also observed from an increase in switching between parties and the decline in selective turnout. Our analysis thus far points to the emergence of a divided party system.

First and second preferences To further understand the changes taking place we examine the relationship between a respondent’s first preference and his or her second preference. While this gives us an idea about party ‘solidarity’, or the extent to which parties matter over individual candidates (see also chapter 7), it also informs us about the nature of this hollowing of the centre. In terms of party solidarity, a voter may express their first and second preferences along party lines or be influenced by specific candidates. Clearly, in some cases, a voter may choose the sole Labour Party candidate followed by the sole Green Party candidate and still maintain that they are voting along party lines. However, for the larger parties, we can see more clearly whether the party label appears to matter. Table 2.5 gives the distribution of second preferences among voters that had an opportunity to express a second preference for another candidate from the same party as their first preference. We can see that they give a second preference for the same party in roughly 62 per cent of cases, rising to 68 per cent in the case of supporters of Sinn Féin. The loyalty of Sinn Féin voters may be somewhat inflated as there are few constituencies where Sinn Féin ran more than one candidate and these areas were particularly favourable to the party. However, considerable party solidarity remains among current supporters of the larger parties. Table 2.5 also shows that many voters express a significant number of second preferences for parties that are similar in outlook. For example, voters giving a first preference to Fine Gael tend to give their next party preference to a

24

The post-crisis Irish voter Table 2.5  Distribution of first and second preferences of voters with an option to give a second preference to the same party   Second preference Fine Gael Fianna Fáil Sinn Féin Labour AAA-PBP Independent Other No other party Sample size

First preference  Fine Gael

Fianna Fáil

Sinn Féin

60 10 1 12 1 8 5 3 991

12 61 6 3 1 10 4 3 650

3 4 68 1 8 10 6 3 241

Source: INES1 2016. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

candidate of its coalition partner, Labour. Voters who express a first preference for Sinn Féin tend to identify candidates of the similarly left-leaning AAA-PBP after they have selected Sinn Féin candidates. To understand the structure more broadly, we must examine all votes while focusing on the pattern of first party preference and the next preferred party preference – their second party preference. Table 2.6 tells us to what extent supporters of one party are willing to support other parties. The larger parties generally receive more second party preferences than smaller parties. This is unsurprising when we consider the fact that that the larger parties are generally more popular and stand more candidates for which nonpartisan rubrics – personal and local factors (see chapter 7) – also apply when choosing to support one candidate over another. That said, their share of second party preferences is generally smaller than their first preference vote share. The relationship between first and second party preferences reveals how the post-crisis voter interprets the party system. There is a close connection between Fine Gael and Labour in the minds of many voters: 33 per cent of voters who gave their first preference to Fine Gael gave their next preference to Labour while 50 per cent of voters who gave their first preferences to Labour gave their next preference to Fine Gael. Similarly, there is a close connection between Sinn Féin and the AAA-PBP: 18 per cent of voters who gave their first preference to Sinn Féin gave their next preference to the AAA-PBP while 31 per cent of voters who gave their first preferences to the AAA-PBP gave their next preference to Sinn Féin. Sinn Féin broadly attracted considerable anti-establishment support.

Table 2.6  Patterns of party choice: Proportion of first preferences for each party going to each possible second next preferred party   First preference party Fine Gael Fianna Fáil Sinn Féin Labour AAA-PBP Green Social Democrats Renua Independent Other parties Total

Next preferred party None

Fine Gael

Fianna Fáil

Sinn Féin

Labour Party

AAA -PBP

Green Party

Social Democrats

Renua Ireland

18 23 18 7 10 4 5 1 14 4 16

0 20 7 50 5 17 21 28 18 5 14

19 0 10 14 4 10 15 22 17 11 12

2 11 0 4 31 5 12 5 13 24 8

33 8 3 0 6 18 10 4 7 4 13

1 3 18 3 0 12 4 3 6 11 5

5 4 4 6 11 0 13 8 3 7 5

3 4 4 5 5 7 0 4 3 4 4

4 6 2 1 3 3 3 0 4 5 3

Source: INES1 2016. Note: Shading used highlights higher values. Sample size: 4,283. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

Independent Other 14 20 28 11 21 22 14 25 14 25 17

1 1 5 0 2 3 4 1 2 0 2

26

The post-crisis Irish voter

Twenty-four per cent of voters who gave their first preference to other parties gave their next preference to Sinn Féin. Independents tended to receive a large number of next preferences from all supporters, particularly Sinn Féin (28 per cent) and the AAA-PBP (21 per cent), while Independent supporters tended to give their next preference to parties according to the overall distribution of first preference votes: in other words, Independent voters look like a cross section of voters for parties. There are thus two ‘systems’ of transfers: some voters voting and transferring their vote between one set of parties, and a second set of voters voting and transferring their vote between another set of parties. This explains the rise in the number of parties being represented without a significant rise in the number of parties that voters are issuing preferences for. The first system of preferences is between Fine Gael, Labour and Fianna Fáil – the established parties towards the right of the left–right spectrum – and the second system is between Sinn Féin and the AAA-PBP towards the left of the left–right spectrum (for more detail, see chapter 3). Some parties receive a much greater proportion or a much lower proportion of next and lower preferences than their first preference share. The Labour Party, for example, had almost twice as many second party preferences as Sinn Féin despite having almost half as many first preferences. Similarly, the Green Party received considerably more second preferences than Renua, despite achieving a similar first preference vote share. These are clear indications of which parties are liked, if not loved, by many (Labour and the Greens) and which parties are loved by some but also disliked by many (Sinn Féin and Renua).

The value of transfers After discussing the information that lower preferences and the ballot offer us, we must also address the overall value of lower preferences. Just 14 per cent of those elected and 4 per cent of all candidates are elected on the basis of their first preference votes, meaning that the remainder, at least in theory, need to attract lower preferences in order to be elected. Arguably, those that polarize opinion may be disadvantaged and do worse in terms of the number of candidates that they get elected given their first preference vote share than, say, those parties that are everyone’s second favourite. The relationship between second preferences and electoral success is conditional on the fact that second and lower preferences from other candidates only matter if preferences actually transfer from one party to another. First preferences only transfer where a candidate is eliminated before the final count or where he or she is elected with a surplus. On the basis of election data from 2016, we can calculate that just 31 per cent of first preference votes were transferred to another candidate. The remaining 69 per cent stayed with the voter’s first preference, who was either elected or eliminated in the final round of counting.

27

Mining the ballot Table 2.7  Party first preference vote shares, percentage of votes transferred from each party and their overall contribution to the share of all transferred votes as taken from the results of the 2016 election (2016 Electoral Data)

  Fine Gael Fianna Fáil Sinn Féin Labour AAA-PBP Green Party Social Democrats Renua Ireland Independent Other parties

First preference vote share

% of party’s first preference vote transferred

Share of all transferred votes

26 24 14 7 4 3 3 2 17 1

13 20 28 48 50 85 46 86 48 96

10 15 12 10 6 7 4 6 26 2

Note: Shading used highlights higher values. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

Table 2.7 shows that a much greater share of the first preference votes of candidates of smaller parties was transferred. While the first preference vote shares of the two largest parties make up half of all first preferences, they only make up a quarter of all transferred votes. By contrast, support for Independent candidates transferred more than that for Fine Gael and Fianna Fáil combined, as the former were much more likely to be eliminated. Furthermore, many of the votes that did transfer from the larger parties transferred internally to other candidates on the party ticket. As such, the value of being the next preferred party among Fine Gael or Fianna Fáil voters was small relative to the size of those parties’ first preference vote. Indeed, in the explicit ‘transfer pact’ between Labour and Fine Gael, the lower preferences from the Labour Party were probably much more valuable than the lower preferences from Fine Gael, despite the former being almost four times the size of the latter. In relation to the aforementioned two systems of transfers, the antiestablishment left-wing one is far more viable than it might initially appear. Although Sinn Féin and the AAA-PBP are more toxic than other parties, they are not significantly disadvantaged. Sinn Féin’s problem of attracting few transfers from Fine Gael is ameliorated by the party’s ability to attract significant transfers from the AAA-PBP, the Social Democrats and other parties that tend to get eliminated and who together account for 12 per cent of all transfers.

28

The post-crisis Irish voter

The impact of attracting lower preferences can be estimated from the relationship between the number of first preferences a candidate receives and whether he or she is elected or not. The reason why this is a probabilistic, rather than deterministic, relationship is due to the impact of lower preferences. A candidate with a given vote share may or may not be elected due to his or her ability to attract a greater or smaller number of lower preference votes and the ability of the party to ‘manage’ the vote between the candidates it is running in a given constituency. For example, at one extreme, Catherine Ardagh of Fianna Fáil failed to get elected in Dublin South Central on 63 per cent of the quota, while at the other end of the spectrum, Independent candidate Katherine Zappone was elected in Dublin South West on 40 per cent of a quota. In the case of Catherine Ardagh, she was overtaken by Bríd Smith of the AAA-PBP who attracted enough transfers from eliminated candidates from Sinn Féin and the Social Democrats. On the other hand, Katherine Zappone was able to overcome a 2,000-vote gap between herself and the Fine Gael candidate by winning sufficient numbers of transfers from eliminated Independent, Green, Labour Party and AAA-PBP candidates. We can approximate the marginal effect of lower preference votes by looking at aberrations in the otherwise assumed relationship between first preference votes and whether the candidate is elected or not. While other factors pertaining to the composition of candidates running in a given constituency will also impact on a candidate’s fortunes, we can assume that these factors will be broadly random from one constituency to the next. Any residual differences between the predicted seats and actual seat count we can attribute to transfers and the internal party management of candidate first preference votes. To estimate the relationship, one must account for differences in the district magnitude (or number of members elected) of constituencies. Constituencies with three, four and five seats require different shares of the vote (the quota) for a candidate to get elected. We therefore estimate the candidate’s first preference totals in terms of the quota that would ensure election. To account for the nonlinear nature of this relationship in 2016 we employ a non-linear model using the square and the cube of the proportion of the quota received, ‘PQ’. This gives us the following formula, which is graphically displayed in Figure 2.2: Probability of winning = logit–1 (–28.6 + 131.6PQ – 212.6PQ2 + 125.6PQ3)

where PQ is the proportion of the quota won in terms of first preference votes. We use this model to estimate how many seats each party would expect to win based on the first preference performance of its individual TDs. We first estimate the probability that each candidate will win a seat. For example, Catherine Ardagh had an 86 per cent chance of winning a seat based on this model while Katherine Zappone had a 13 per cent chance of winning. We

Mining the ballot

29

Figure 2.2  The estimated relationship between the quota received and the probability of winning a seat. Source: ‘ElectionsIreland’, http://www.electionsireland.org 1998–2017, Christopher Took and Seán Donnelly.

can then estimate how many seats the party should expect to win in order to understand the impact of transfers on overall seat numbers. Table 2.8 compares the number of seats won by each party with the estimated number of seats the party would have won if transfers had been distributed according to the model. What Table 2.8 shows is that the effect of transfers in 2016 was minimal. Only a few seats were affected positively or negatively by lower preferences. The toxicity of Sinn Féin cost the party one seat, whereas the transfer-friendliness of the Greens produced a single additional seat over what would otherwise have been expected. Independent and AAA-PBP candidates also performed better than what would have been otherwise expected, perhaps on account of significant numbers of lower preferences transferring between them.

Conclusion Our analysis of ballots (first preferences and lower preferences) reveals several important features of the post-crisis Irish voter. Firstly, the unprecedented changes in first preference voting have been underpinned by enormous amounts of party switching rather than by selective turnout – a behavioural trait which appears to have declined in the aftermath of the crisis. This change reflects the erosion of long-term party allegiances. In previous elections, party allegiances may have remained strong where loyalists stayed at home rather than voted for another party. Post-crisis, as this comparison suggests, larger numbers of

30

The post-crisis Irish voter

Table 2.8  Number of seats won, number of occasions the party would have won without any additional counting, modelled estimates of expected seats and effective seat bonus due to vote management and transfers Party Fine Gael Fianna Fáil Sinn Féin Labour AAA-PBP Social Democrats Green Renua Independent, Others

Actual seats 49 44 23 7 6 3 2 0 23

First preference rank wins 51 46 23 6 5 3 1 0 22

Model expected seats 50.6 44.2 23.8 7.6 5.3 3.5 1.1 1.1 19.8

Rounded model seat bonus −2 0 −1 −1 1 0 1 −1 3

Note: Based on election count data from the 2016 general election. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

people are actively switching rather than staying at home. Finally, those that stay at home are becoming more consistent in that form of voting behaviour. Secondly, there is significant evidence of some hollowing of the centre ground in Irish politics or, in other terms, of the emergence of an antiestablishment left wing in Sinn Féin and the AAA-PBP. While a number of new parties have emerged, the number of party preferences expressed by voters has barely changed. This is the result of the emergence of two separate party systems with preferences transferring between establishment parties towards the right of the left–right spectrum and preferences transferring between nonestablished parties towards the left of the left–right spectrum. While the group of non-established parties have little appeal among voters of established parties, they are not disadvantaged as they benefit from the relatively larger amount of lower preferences available from the elimination of smaller non-established candidates and parties. This means that in spite of attracting a relatively low number of first preferences, non-established parties are barely penalized by the fact that the electoral system also gives weight to lower preferences. After two elections it is as of yet unclear what direction the Irish political system is likely to take. The party system could very well revert to type, with Sinn Féin coming in from the cold attracting and relying on transfers from the established parties, or a bipolar system could emerge around Sinn Féin and Fine Gael. Either way, at this point it seems unlikely that the level of volatility in the two most recent elections can be sustained.

Mining the ballot

31

Notes 1 Note that those who could not remember how they voted were excluded from the analysis. The data are fitted iteratively so that the table marginals reflect 1997 and 2002 results. The shade of each cell relates to the relative magnitude of the data, with higher values highlighted in a darker colour. 2 One caveat in respect of applying an analysis of reported turnout is the quality of the electoral register, which is larger than the number of valid electors on account of a significant numbers of duplicates and deceased people that remain on the register.

References Cunningham, Kevin, and Michael Marsh. 2018. ‘Voting behaviour’, in John Coakley and Michael Gallagher (eds), Politics in the Republic of Ireland. 6th edition. London: Routledge, pp. 137–63. Elkink, Johan A., and David M. Farrell (eds). 2016. The Act of Voting: Identities, Institutions and Locale. London: Routledge. Farrell, David M., and Jane Suiter. 2016. ‘The election in context’, in Michael Gallagher and Michael Marsh (eds), How Ireland Voted 2016: The Election that Nobody Won. Basingstoke: Palgrave Macmillan, pp. 277–92. Gallagher, Michael. 1978. ‘Party solidarity, exclusivity and inter-party relationships 1922–1977: The evidence of transfers’, Economic and Social Review. 10(1): 1. Gallagher, Michael. 1992. ‘The election of the 27th Dáil’, in Michael Gallagher and Michael Laver (eds), How Ireland Voted 1992, Dublin: Folens/PSAI Press, pp. 57–78. Gallagher, Michael. 1999. ‘The results analysed’, in Michael Marsh and Paul  Mitchell (eds), How Ireland Voted 1997, Boulder: Westview Press, pp. 121–50. Gallagher, Michael. 2003. ‘Stability and turmoil: Analysis of the results’, in Michael Gallagher, Michael Marsh and Paul Mitchell (eds), How Ireland Voted 2002. London: Palgrave, pp. 88–118. Hix, Simon, and Michael Marsh. 2011. ‘Second-order effects plus pan-European political swings: An analysis of European Parliament elections across time’, Electoral Studies. 30: 4–15. Mair, Peter. 2011. ‘The election in context’, in Michael Gallagher and Michael Marsh (eds), How Ireland Voted 2011. London: Palgrave Macmillan, pp. 283–97. Marsh, Michael. 1998. ‘Testing the second-order election model after four European elections’, British Journal of Political Science. 28: 591–607. Marsh, Michael, Richard Sinnott, John Garry and Fiachra Kennedy. 2008. The Irish Voter: The Nature of Electoral Competition in the Republic of Ireland. Manchester: Manchester University Press.

32

The post-crisis Irish voter

O’Kelly, Michael. 2016. ‘Locality in Irish voter preferences’, in Johan Elkink and David Farrell (eds), The Act of Voting: Identities, Institutions and Locale. London: Routledge, pp. 137–60. Sinnott, Richard. 1995. Irish Voters Decide: Voting Behaviour in Elections and Referendums since 1918. Manchester: Manchester University Press. Sinnott, Richard, and James McBride. 2011. ‘Preference voting under PR-STV, 1948–2011’, in Michael Gallagher and Michael Marsh (eds), How Ireland Voted 2011. London: Palgrave Macmillan, pp. 205–21.

3 Ideological dimensions in the 2016 elections Kevin Cunningham and Johan A. Elkink

Introduction This chapter explores the emergence of new and old ideological dimensions in the context of a changing party system. The Irish voter is characterized as non-ideological: prone to issuing voting preferences based on locality (Chubb, 1957; Carty, 1981; Marsh, 2007) and family traditions (Carty, 1981; Sinnott, 1995; Gallagher and Marsh, 2002). The political system has thus been accurately marked as sui generis (Whyte, 1974), an ‘exception’ (Carty, 1981), and even a ‘problem child’ (Urwin and Eliassen, 1975) as it eschews the traditional left– right basis of political competition in Western European politics. This has been most clearly illustrated by the weak influence of social class and of left-wing politics. It is manifested in the dominance of two ideologically indistinguishable catch-all parties: Fianna Fáil and Fine Gael (Benoit and Laver, 2003; Lutz, 2003) and the persistent weak levels of support for the Labour Party, the long-term flag-bearer of the left or, more precisely, centre-left. Irish society has been changing quite dramatically. Party identities have declined, the Irish political system entered a period of very considerable volatility following the global financial crisis in the early 2010s and new dimensions of political competition defined by globalization seem to be all but replacing traditional economic left and right in Western Europe. In this context, this chapter re-examines the ideological nature of political competition in Ireland. We first discuss how and why ideology may have a greater role in defining party competition both historically and in the immediate context of postcrisis Ireland. Second, we discuss the key dimensions of ideology, new and old, before exploring the extent to which attitudinal questions underpinning those dimensions reflect the preferences of voters. This helps us to answer questions such as whether ideology is a significant component of party choice and whether voting behaviour in Ireland ties in with the thesis that contemporary European politics revolves around a new cleavage of ‘winners’ and ‘losers’ of globalization

34

The post-crisis Irish voter

as well as other cleavages around left-wing populism potentially more relevant to the post-crisis voter and the politics of austerity. The evidence suggests that classic left–right self-placement is a significant predictor of vote choice dividing Fine Gael and Fianna Fáil from Sinn Féin and Labour. We find that this is related to two underlying dimensions: economic left–right issues divide Fine Gael from Sinn Féin, and socio-moral issues divide Fianna Fáil from Labour. We understand that these two features explain left and right. A second dimension relating to populism, the importance of local issues and evaluations of the economy in the past five years, divides Fine Gael and Labour from Fianna Fáil and Sinn Féin. Due to the significant impact of left and right, we explore this in detail to find that voters position themselves close to the ideological positions of their chosen party. Voters position Fianna Fáil and Fine Gael on the right, Labour in the middle and Sinn Féin on the left. Finally, we use the sequence of preferences on replicated ballots from the RTÉ/ INES exit poll (INES1) to understand the dimensions of political competition as derived from actual voting behaviour. This dimensional analysis suggests that preferential ordering reflects the aforementioned dimensions with elements of economic left–right, socio-moral left–right, populist versus elitist and government versus opposition underpinning the relative positions of political parties in the two-dimensional space. Typical issues relating to globalization more specifically are not yet a feature of Irish politics but it is clear that there is some overlap between left–right and the populist-elitist dimensions in this regard.

Decline of environmental constraints Lipset and Rokkan’s (1967) analysis of party systems suggests that Ireland’s apparent uniqueness is due to fundamental differences between Ireland and other Western European countries. The political system that emerged in 1922 inevitably reflected the divisions of the deeply divisive civil war between radical nationalists and moderate nationalists. Only these two groups had the organizational capacity to establish a viable political organization and both parties sought to inherit the political capital of their ancestor party Sinn Féin. Environmental factors also inhibited the emergence of a classic left–right divide, notably an agrarian economy and a dominant Catholic Church. Little challenged by a cautious Labour Party (Gallagher, 1985: 92), Fianna Fáil filled much of the political space that would otherwise be held by a party defined on the left: the populist stances of Fianna Fáil and its capacity to appeal to rankand-file trade union members further undermined any significant support base for the left. In the intervening period, as European democracies moved from manufacturing-oriented economies to service-oriented economies, Ireland

Ideological dimensions in the 2016 elections

35

moved from its agrarian-oriented economy to a service-oriented economy. The role of the Catholic Church in political life has subsided considerably and the status of Northern Ireland following the Good Friday Agreement served to de-politicize the historical fault line between the two civil war parties. These changes mean that political debates in Ireland increasingly reflect typical debates in other European countries. Evidence of statistically significant differences between party supporters emerged from the 1980s onwards: on left–right positions, religious-moral outlook, nationalism and environmentalism (Mair, 1986; Laver, Mair and Sinnott, 1987; Breen and Whelan, 1994; Farrell, 1999; Benoit and Laver, 2003). In highly politicized referendums on abortion and divorce, Fianna Fáil adopted the more socially conservative position, opposing the introduction of divorce and supporting stricter legislation in relation to abortion; Labour typically adopted the more liberal stance. Expert studies suggest that the Irish political system could be characterized in terms of two dimensions: social policy and tax-spend policy, with the two larger parties in one corner representing the economically right and socially conservative positions and Labour and all the smaller left-leaning parties in the opposite corner (Laver, 1992). However, although statistically significant, the substantive impact of this has been weak (Marsh et al., 2008). Ideology on the socio-moral axis may significantly influence support for the Labour Party, but the party system and the relative levels of party support have barely changed. Only a small subset of voters had perhaps interpreted these differences as a basis for determining which party they would vote for. This is mainly due to the persistent dominance of – and insignificant differences between – Fianna Fáil and Fine Gael. For ideological dimensions to explain party support, Fianna Fáil and Fine Gael would need to lose a significant amount of their vote. It is not so remarkable that the two parties would dominate political competition in the nascent state, but what is remarkable is that the same dynamic would persist many years after the aforementioned unique environmental conditions had subsided. As party identification – required to sustain this order – declined steadily from the 1980s (Marsh, 2006), a change event such as the post-2008 economic crisis was necessary for a significant redistribution of party preferences outside of the two larger parties. The elections of 2011 and 2016 were among the most volatile of all elections in Western Europe since 1945, the 2011 election earning the title of the most volatile election without a significant new entrant (Mair, 2011). The two parties averaged a total of 79 per cent of the vote in the 1980s, 65 per cent in the 1990s and 66 per cent in the 2000s. Following the economic crisis, support for the two parties declined further to a 52 per cent average in the 2010s with left and left-of-centre parties averaging a total of 32 per cent.

36

The post-crisis Irish voter

Ideological dimensions Left and right The 2008 global financial crisis had a profound and devastating impact on the Irish economy and much of the blame was placed on Fianna Fáil. In the ‘earthquake’ election of 2011, support for Fianna Fáil collapsed. For the first time since 1927 it was not the largest party and for the first time ever it was the third largest party. It was this collapse that precipitated some belief that a new left–right cleavage might emerge in Irish politics (for discussion, see Marsh, Farrell and McElroy, 2017). The party system had, in Fine Gael, a classic centre-right party with significant support, but it required the emergence of something from the left to develop a classic left–right basis of political competition. In 2011 the left performed better than in any previous election, winning 31 per cent of the vote. For the second time, the Labour party had won 19 per cent of the vote. However, the party’s decision to go into government in coalition with Fine Gael greatly undermined the development of a left–right basis of political competition (Marsh and Cunningham, 2011) as, following a longstanding pattern, in the 2016 election support for the minor coalition partner, Labour, collapsed. The 2016 election saw the party system take another turn with the partial recovery of Fianna Fáil and the continuing growth of Sinn Féin. Today’s main left-wing opposition is Sinn Féin, and while the party’s roots look more like the civil war politics of old with a strong association with Northern Ireland and constitutional nationalism, the party’s support is much more clearly defined in terms of a working-class social base than perhaps any other party before it (see chapter 4). The rise of Sinn Féin in addition to the Labour Party, and a multitude of smaller left-leaning forces (including the Anti-Austerity Alliance, People Before Profit (AAA-PBP), Social Democrats, Socialist Party and Workers’ Party) meant that support for the left in total remained high in 2016 at 29 per cent. Data from the 2016 general election would suggest that ideology was important for some voters and unimportant for others. In the post-election survey (INES3) voters were asked on a scale of 0 to 10 how important it was that the candidate they chose had the same political viewpoint as the voters had themselves. The average score for this question was 6.3, with 39 per cent of Irish voters recording 5 or less on the scale. To understand the magnitude of this we can compare it with other factors that might be considered important, such as their chosen candidate having the same age, the same gender, the same level of educational attainment, the same social class, or being from the same area as the voter. Political viewpoint is the most important consideration but it does not dominate. In 31 per cent of cases the candidate’s political viewpoint was the most important consideration and in a further 22 per cent it was at least as important as the next most important factor. However, for 47 per cent of voters

Ideological dimensions in the 2016 elections

37

a factor other than political viewpoint was more important. In 29 per cent of cases, having the same level of education was of higher importance. In 26 per cent of cases, where the candidate lived was more important than political viewpoint and in 20 per cent of cases the candidate being from the same social class was more important. A theme emerges where, for a significant number of voters, the local and personal relationships they have with candidates are deemed to be more important than their ideological positions (see chapter 10 for a more extended discussion). On their own admission, ideology only matters much for roughly half of all voters and there is a pattern to this. When we compare this result against the person’s left–right position a clear trend appears. Among ‘left-leaning’ voters (those that place themselves as 4 or lower on a left–right scale), 42 per cent rate a candidate’s political viewpoint as more important than any other issue and, for 23 per cent of them, it was at least as important as the next most important factor. Among the most left-leaning voters (defining themselves as 2 or lower on the left–right scale), 77 per cent regard the candidate’s ideological viewpoint as at least as important as any other consideration. By contrast, on the righthand end of the spectrum, there is no such sharp focus on political viewpoint. It is notable that political viewpoint is considerably less important for supporters of Fianna Fáil (for just 29 per cent it was the most important) and Independent candidates (for 26 per cent it was the most important). This compared with supporters of Labour (35 per cent), the AAA-PBP (51 per cent) and the Green Party (58 per cent). Clearly, ideology is important, but seemingly it is more important for left-leaning voters intending to vote for left-wing parties.

Local factors Structural factors also complicate the development of a clear left–right divide in Irish politics. One outcome of Ireland’s single transferable vote (STV) electoral system is that it allows support for Independent candidates based on local values and patronage rather than left or right ideologies. These candidates benefitted enormously from the collapse of Fianna Fáil as they were uniquely positioned to capture some of the anti-system populist vote that might in other jurisdictions be captured by a left-wing political party (Carty, 1981; Komito and Gallagher, 2010; Marsh, 2007; Bowler and Farrell, 2017) (see chapter 7 for more details). The exit poll (INES1) asked respondents whether the party or the candidate was more important. Excluding those that did not know, 46 per cent stated that the party was more important than the candidate whereas 54 per cent stated that the candidate was more important than the party. Arguably, as candidates can be supported on policy grounds, it is almost certainly the case that national policy claims made by candidates are not definitive. Furthermore, when we ask if they would still vote for the same candidate if they ran for another party, 40 per cent say they would, and 17 per cent state that it would depend on which party.

38

The post-crisis Irish voter

Globalization The local-national division is reflected in new dimensions of political competition. Across Europe, party choice is typically structured by a single left–right ideological dimension (e.g. Oppenhuis, 1995; Van der Eijk and Franklin, 1996). However, this is deemed to be insufficient in the context of the pressures of increasing globalization (Kriesi et al., 2006; 2008). Globalization has yielded winners and losers as increasing international integration and the role of market forces on economies have reduced government influence on local economies. As a result, solutions designed in terms of traditional left-wing or right-wing political responses have become much less effective. This has resulted in support for populist parties that have attempted to address rising frustration and distrust of politicians with support for protectionist solutions seeking to reverse globalization. The behaviour of political parties is now defined by a socio-economic as well as a socio-cultural dimension. Cross-cutting traditional economic left– right divisions and the conservative-liberal dimension are sometimes explained in terms of the ‘winners’ and ‘losers’ of globalization (Kriesi et al., 2006; 2008), or, equivalently, in terms of those that see themselves as ‘anywhere’ as opposed to those that have rooted themselves ‘somewhere’ (Goodhardt, 2017; Jennings and Stoker, 2017). The ‘winners’ or ‘anywheres’ may be characterized as having higher levels of education, being economically mobile and valuing openness and autonomy. They consist of a combination of middle-class left-wing voters with cosmopolitan values and liberal right-wing voters favouring international liberalization and free trade. The ‘losers’ or ‘somewheres’ are more often rooted, particularly in towns and rural areas; they value their local identity, group attachments, community and familiarity. They combine working-class voters with traditional left-wing values and those resorting to a stronger antiimmigration, more right-wing perspective (Teney, Lacewell and De Wilde, 2014). This dimension relates closely to the expected impact of globalization on low-skilled and high-skilled workers, based on economic theory around factor endowments, which predicts that in rich countries, lower-skilled workers will suffer more negative competition from immigration (Mayda and Rodrik, 2005; O’Rourke and Sinnott, 2006). In many cases, this globalization dimension is underpinned by a populist dimension that juxtaposes ‘the pure people’ against ‘the corrupt elite’ (Mudde, 2004). This is most evident in Ireland in relation to the emergence of the antiwater charges protests, in line with what Meyer and Tarrow (1998) termed ‘movement societies’ where political protest has become an integral part of Western European countries’ political life, employed with greater frequency by more diverse constituencies to represent a wider range of claims than ever before. We arrive at several competing or complementary dimensions of voting behaviour: economic left–right, socio-moral left–right, globalization,

Ideological dimensions in the 2016 elections

39

populism, the value of the local candidate and the constitutional status of Northern Ireland.

Analysis of ideological dimensions We explore each ideological dimension by first seeking to understand the extent to which it relates to party preferences. For parsimony, we identified specific attitudinal questions that represented the dimensions most closely. With respect to the traditional economic left and right, we look at both tax and spend and income inequality attitudes. In relation to the socio-moral dimension we use the respondents’ attitudes towards abortion. Respondents record their own position on abortion from a total ban on abortion, 0, to a complete liberalization, 10. Those that give themselves a score between 0 and 3 are classified as being ‘prolife’, those that give themselves a score between 7 and 10 are classified as being ‘pro-choice’ and the remainder are classified as being ‘unsure’. Respondents are also asked about their position on a ‘left–right scale’. Those with a score between 0 and 3 are classified as being ‘left’, those with a score between 7 and 10 are classified as ‘right’ and those in between are classified as being of the ‘centre’. We explore the new globalization dimension via responses to an attitudinal question on immigration. The role of constitutional nationalism in terms of the status of Northern Ireland ties in with this. In relation to populism, we looked at attitudes towards the people and elites. The role of a TD in terms of providing a local service relates to the populist and globalization dimensions. Because many of these dimensions overlap we estimated the relationship between these attitudes and vote choice independently. Specifically, we modelled the relationship between each attitudinal variable and each respondent’s party choice using a multi-level regression model, having reformatted the data so that each case represented a respondent-party choice. Each model enabled us to estimate each party’s support in each category. The mean support of the party and the support within the subgroup of each set of attitudes are given in Figures 3.1a to 3.1d. The respondent’s position on the left–right scale is significant with respect to the Fianna Fáil, Fine Gael, Sinn Féin and AAA-PBP votes. The Fine Gael vote rose from 24 per cent to 38 per cent among those who defined themselves as being on the right. Similarly, the Fianna Fáil vote increased from 21 per cent to 33 per cent. In relation to the left, Sinn Féin’s support increased from 17 per cent to 36 per cent whereas AAA-PBP increased their vote from 5 per cent to 13 per cent among those that identified themselves as left. In the case of Fine Gael, this appears to be related to significant (if not substantially) right-wing positions on economic and socio-moral grounds. The party performs 5 per cent better among those who disagree with increases in tax and spend policies, 4 per cent better among those who would not like the state to act to reduce

40

The post-crisis Irish voter

income inequality and 3 per cent better among those who adopt a position that might be classified as being ‘pro-life’. The factors that underpin Fianna Fáil as a right-wing party reflect socio-moral themes more than economic themes. Fianna Fáil support increases by 4 per cent to 25 per cent among those that identify themselves as ‘pro-life’. Continuing with the socio-moral themes, the Labour Party performs 2 per cent better among those that are pro-choice. Globalization issues seem to map closely onto the socio-moral dimension, but weakly so. Both Fine Gael and Fianna Fáil perform 2 per cent better among those who would like to see greater limits on non-EU immigration; the Green Party performs 5 per cent better among those who would disagree with this. Immigration has a limited impact on Irish politics: nationalist sentiment is more commonly reflected in relation to Northern Ireland, for which Fine Gael and Fianna Fáil hold a much less nationalist position than Sinn Féin, who perform 10 per cent better among those who would like to see Northern Ireland unite with the Republic of Ireland. The issue of Northern Ireland appears to be a core belief for Sinn Féin supporters rather than a defining feature of the political system. The populism dimension is better defined. Fine Gael performs 7 per cent better among those who would disagree that ‘people, rather than politicians should make decisions’. Fianna Fáil and Sinn Féin are more likely to agree with the alternative, more populist view of people rather than politicians making decisions as both parties perform 2 per cent better among those who agree. Equivalently, Fianna Fáil and Independents attract 3 and 4 per cent more support, respectively, among those who agree that TDs should provide a local service, whereas Fine Gael and Labour perform 4 and 5 per cent better among those who would disagree with the idea that TDs should provide a local service. We may characterize the party system as being defined as right versus left, in economic terms dividing Fine Gael from Sinn Féin and in socio-moral terms dividing Fianna Fáil from Labour. Another dimension cuts across this, dividing the more populist and localist Fianna Fáil and Sinn Féin from Fine Gael and Labour. Immigration and the role of Northern Ireland are much weaker. Even if statistically significant in some cases, the marginal impact of these variables renders them somewhat irrelevant. To explore this in greater detail we examine the extent to which each ideological dimension relates to support for each party while accounting for other ideological dimensions. We conduct a series of separate regression analyses looking at the relationship between attitudinal responses and the extent to which respondents ‘like’ each party. This allows us to understand the relationship between dimensions – whether some dimensions add explanatory power and whether others are merely related to one another. Note that we must rely on a single dataset (INES2) for this form of analysis, limiting the range of dimensions that we can assess. We therefore rely on the dataset that has adequate measures of left and right, socio-moral left–right (through religion),

Ideological dimensions in the 2016 elections

41

Figure 3.1a Mean support for Fine Gael and predicted support for Fine Gael within attitudinal categories. Source: INES1 2016, INES2 2016 and INES3 2016.

Figure 3.1b  Mean support for Fianna Fáil and predicted support for Fianna Fáil within attitudinal categories. Source: INES1 2016, INES2 2016 and INES3 2016.

42

The post-crisis Irish voter

Figure 3.1c  Mean support for Sinn Féin and predicted support for Sinn Féin within attitudinal categories. Source: INES1 2016, INES2 2016 and INES3 2016.

Figure 3.1d  Mean support for Labour and predicted support for Labour within attitudinal categories. Source: INES1 2016, INES2 2016 and INES3 2016.

Ideological dimensions in the 2016 elections

43

economic left–right (through tax and spend attitudes), populism (through attitudes towards people or politicians making decisions) and globalization (through attitudes towards immigrants). We do not have an adequate measure to replicate the Northern Ireland dimension and in relation to socio-moral issues, we rely on membership of a religion. Although this is not the question on abortion, we do know that these are strongly related to one another. As per INES1, 50 per cent of those that are atheist, agnostic or spiritual advocated a very pro-choice position compared with just 23 per cent of those who had a religion. Seventy-seven per cent of those that were atheist, agnostic or spiritual adopted a pro-choice position compared with 50 per cent of those with a religion. We also control for demographic variables and perceptions of the economy which relate to support for parties in this specific election. We control for age, gender, social class, education, work status and the type of area voters live in. We also control for their perception of the performance of the economy – whether it got worse, stayed the same or got better (see also chapter 6 in this volume). This is an important control for there may very well be significant numbers that momentarily dislike their typical party of choice due to their perceptions of the economy. The results of the regressions are given in Table 3.1. Our first observation is that a person’s left–right position is a highly significant predictor of which parties a voter is likely to favour. Those that are more left-wing are also more likely to like Sinn Féin, whereas those that are more right-wing are more likely to like Fine Gael or Fianna Fáil. The second most prominent feature is the populist dimension, that is, the extent to which people, not politicians, should make decisions. Those that like Fine Gael and Labour are much less likely to adhere to this view whereas those that tend to like Sinn Féin are more likely to agree with it. This appears to be more prominent once the respondent’s left–right position is accounted for. We also sought to determine the true difference between supporting Fianna Fáil over Fine Gael as estimated by subtracting the self-reported extent to which voters liked Fianna Fáil from the self-reported extent to which they liked Fine Gael. Those that tend to like Fianna Fáil more than Fine Gael are also more likely to adopt the more populist perspective. This corroborates earlier evidence supporting a populist dimension with Labour and Fine Gael towards the elitist end and Fianna Fáil and Sinn Féin towards the populist end. The regression suggests that Fianna Fáil supporters are more likely to be in favour of tax and spend policies. One should bear in mind that these effects are estimated while already accounting for the party’s left–right position. The party is perhaps perceived as being right wing on the basis of its socio-moral position rather than economic issues. Certainly, those that like Fianna Fáil are significantly more likely to adhere to a religion while those that like Labour are significantly less likely. Our analysis of this variable supports earlier evidence

Table 3.1  Regression analysis estimating the extent to which a respondent likes Fine Gael, Fianna Fáil, Sinn Féin, Labour, and the difference in the extent to which they like Fianna Fáil over Fine Gael    

Fine Gael

 

Fianna Fáil

 

Sinn Féin

 

Labour

 

Estimate P-Value

 

Estimate P-Value

 

Estimate P-Value

 

Estimate P-Value

 

Intercept 0.171 Left −0.521 Right 0.380 Agree people should −0.290 make decisions Unsure people should −0.014 make decisions In favour of tax and 0.041 spend Unsure of tax and 0.071 spend Have religion 0.172 Agree culture harmed 0.107 by immigrants Unsure culture harmed 0.087 by immigrants

0.718 0.017 0.022 0.069

0.257 −0.090 0.062 −0.019

0.000 *** 0.001 *** 0.003 ** 0.355

0.259 0.121 −0.086 0.091

0.000 0.000 0.000 0.000

0.963

−0.093

0.013 *

−0.019

0.890

0.124

0.001 ***

0.816

0.096

0.394 0.611 0.730

* * .

 

Estimate P-Value  

0.560 −0.028 −0.023 −0.056

0.000 *** −0.290 0.245 0.026 0.224 −0.025 0.002 ** 0.048

0.000 *** 0.470 0.377 0.078 .

0.637

−0.016

0.632

−0.083

0.095 .

−0.009

0.811

0.056

0.088 .

0.120

0.016 *

0.013 *

−0.050

0.224

0.028

0.406

0.084

0.102

0.134 0.021

0.000 *** 0.430

0.048 −0.008

0.083 . 0.785

−0.068 −0.047

0.003 ** 0.044 *

0.092 0.001

0.008 ** 0.982

0.046

0.152

0.826

−0.004

0.891

0.027

0.529

0.008

*** *** *** ***

Fianna Fáil–Fine Gael

Economy was same Economy got worse University degree Working class – C2DE Farmers Female Homemaker Part-time Retired Student Unemployed Large town or city Rural Small town Age 25–34 Age 35–44 Age 45–54 Age 55–64 Age 65+ Source: INES2 2016.

−0.769 −1.444 0.153 −0.287 0.331 0.229 −0.102 −0.034 0.329 −0.139 −0.480 −0.072 0.199 0.099 0.077 −0.179 −0.155 −0.146 −0.160

0.000 0.000 0.354 0.114 0.322 0.159 0.745 0.883 0.309 0.729 0.151 0.713 0.340 0.719 0.823 0.606 0.662 0.696 0.713

*** 0.002 *** −0.059 0.051 −0.044 0.000 −0.039 0.006 0.002 0.007 −0.048 −0.086 0.026 0.066 0.043 0.001 0.018 0.006 0.022   0.082

0.919 0.150 0.016 0.055 0.996 0.057 0.879 0.936 0.870 0.339 0.037 0.290 0.012 0.216 0.988 0.681 0.901 0.645 0.134

  * . .

* *

 

0.071 0.103 −0.051 0.103 −0.089 −0.102 0.020 0.002 −0.013 0.042 0.007 0.056 0.012 0.067 0.016 −0.018 −0.046 −0.057 −0.092

0.004 0.018 0.022 0.000 0.042 0.000 0.638 0.938 0.766 0.435 0.867 0.037 0.668 0.073 0.736 0.701 0.340 0.257 0.116

** * * *** * ***

* .

 

−0.131 −0.238 0.023 −0.076 −0.018 0.080 −0.027 −0.023 0.061 −0.025 0.006 0.004 0.002 −0.006 −0.074 −0.081 −0.048 −0.087 −0.102

0.000 0.000 0.219 0.000 0.615 0.000 0.431 0.379 0.094 0.576 0.869 0.870 0.931 0.839 0.058 0.038 0.231 0.037 0.037

*** ***

0.181 0.254 0.017 *** 0.021 −0.072 *** −0.090 0.024 0.012 . −0.067 −0.015 0.013 0.043 0.022 0.019 . −0.018 * 0.057 0.040 * 0.054 * 0.119

0.000 0.000 0.531 0.490 0.186 0.001 0.644 0.765 0.213 0.819 0.804 0.193 0.527 0.678 0.755 0.322 0.501 0.384 0.101

*** ***

**

 

46

The post-crisis Irish voter

indicating that the socio-moral dimension was one which had Labour at one end and Fianna Fáil at the other. The controls account for perceptions of economic performance, with significant differences between those that support the government (Fine Gael and Labour) over the opposition (Fianna Fáil and Sinn Féin), and demographics, with Sinn Féin supporters in particular more likely to be working-class, male and without a university degree.

Left and right in Ireland Our analysis of ideology leads to a recognition of the importance of understanding what left and right means to voters in Ireland. Policies that define the left or right can vary over time and across countries. Although voters may very well identify themselves as being on the left or the right it is not immediately clear what this means. The left and right may be characterized in terms of economic policy (for income redistribution or tax and spend policies) or social policy, reflecting the established position of the church set against more liberal social values, or indeed newer dimensions around globalization. Using a regression analysis, we can estimate the left and right positions of voters in terms of attitudinal responses to key issues, while controlling for demography and economic voting behaviour. On the basis of the results of this regression analysis, given in Table 3.1, we observe that left–right is related to traditional economic and socio-moral issues. Although left–right does not appear to be significantly related to the populist dimension, it is strongly related to both economic tax and spend policies as well as to the socio-moral agenda. Those that are more right-wing are significantly less likely to be in favour of increases in tax and spend policies. In terms of the socio-moral agenda, those that pronounce that they are atheist, agnostic or labelled as someone without a religion tend to be more likely to position themselves on the left, but only marginally so. Those that are more likely to believe that Ireland’s culture is harmed by immigrants are also marginally more likely to position themselves towards the right. The model explains 15 per cent of the variance in left–right position. Although these figures remain relatively low, they represent an increase from 4 per cent for the pre-crisis voter (Marsh, 2007). To understand left and right further in an Irish context we use the voters’ interpretations of where each party sits on the left and right scale. If there were no left or right in Irish politics, we would expect the party system to be clustered in the centre of the political space. Figure 3.2a is a set of box plots depicting where Irish voters placed Irish parties. For the purposes of comparison, Figure 3.2b is a set of box plots of where British voters placed British parties (using data from the 2015 British Election Study). The two plots help us to compare the parties within the Irish system and against a classic left–right political system as in British politics.

47

Ideological dimensions in the 2016 elections Fine Gael

UKIP

Fianna Fáil Conservative

Party

Green Lib Dem

Labour Soc Dem

Labour

Sinn Féin Green AAA-PBP Self-Placement

Self-Placement 0

1

2

3

4

5

6

7

8

9

Left Right (Irish voters on Irish Parties)

10

0

1

2

3

4

5

6

7

8

9

10

Left Right (British voters on British Parties)

Figure 3.2a and 3.2b  Irish and UK voter’s self-placement and placement of political parties. Source: INES2 2016, BES2015.

Firstly, there is little overlap between Fianna Fáil-Fine Gael and the parties on the left: Labour, Social Democrats, Sinn Féin and AAA-PBP. Coupled with the fact that these latter parties make up almost 30 per cent of the vote, it is clear that the voters understand the positions of the parties in terms of left and right. Fianna Fáil and Fine Gael occupy much of the centre-right where, unlike the British voter, most Irish voters are positioned. Our comparison between the Irish party system and the British party system reveals that there are only subtle differences. The Irish voter is a little to the right of the British voter and the Irish party system is a little to the left of the British party system, if a little more concentrated. Fianna Fáil and Fine Gael are both much closer to the centre of the scale than the Conservatives and UKIP. Irish Labour, the Irish Green Party and the Social Democrats are to the left of the Liberal Democrats. Finally, Sinn Féin and the AAA-PBP are marginally to the left of the British Labour Party and the British Greens. The difference between Sinn Féin and Fine Gael is comparable to the difference between the British Labour Party and the Conservatives. To further understand how far left and right influences party choice we analyse the relationship between the self-placements of voters and placements of their chosen party. As Table 3.2 shows, the self-placement of supporters is close not only to their assessment of their chosen party but also to the assessment of their chosen party by non-supporters. It is noticeable that, aside from supporters of Sinn Féin, supporters of parties tended to position their

48

The post-crisis Irish voter

Table 3.2  Regression analysis of estimated left–right position  

Estimate

P-value

Intercept Agree people should make decisions Unsure people should make decisions In favour of tax and spend Unsure of tax and spend Have religion Agree culture harmed by immigrants Unsure culture harmed by immigrants Economy was same Economy got worse University degree Working class – C2DE Farmers Female Homemaker Part-time Retired Student Unemployed Large town or city Rural Small town Age 25−34 Age 35−44 Age 45−54 Age 55−64 Age 65+

5.116 −0.197 −0.311 −1.027 −0.784 1.264 0.444 0.559 −0.424 −0.345 0.029 0.029 0.561 −0.255 −0.096 −0.240 0.026 0.077 0.019 −0.140 −0.065 0.071 0.257 0.390 0.468 0.594 1.137

0.000 0.191 0.259 0.000 0.006 0.000 0.022 0.018 0.012 0.250 0.850 0.863 0.065 0.094 0.741 0.265 0.931 0.837 0.950 0.446 0.737 0.783 0.426 0.228 0.156 0.088 0.005

  ***

*** ** *** * * *

. .

. **

Source: INES2 2016.

party further away from the centre of political competition than they placed themselves. For example, the average Fine Gael supporter places him or herself at 6.3, and his or her party at 7.2, and non-supporters place Fine Gael at 6.5. This suggests that voters place their chosen party at a position more extreme than their own position, in line with the policy discounting model advocated by Grofman (1985). It is also interesting to note that supporters of established, mainstream parties – Fianna Fáil, Fine Gael, Labour and the Green Party – tended to place these parties further from the centre than non-supporters of each of these parties. However, supporters of non-established parties – Sinn Féin and AAA-PBP – tended to place those parties closer to the centre than non-supporters of these parties.

49

Ideological dimensions in the 2016 elections

Table 3.3  Average assessments of supporters, non-supporters and the average position of supporters themselves Party Fine Gael Fianna Fáil Sinn Féin Labour Green Party Social Democrats AAA-PBP

Supporter’s Non-supporter’s average Supporter’s average average assessment assessment left–right self-placement 7.2 6.8 4.7 4.1 4.0 3.9

6.5 6.4 2.7 4.3 4.5 3.8

6.3 6.3 4.4 5.1 4.4 4.3

2.8

2.3

3.8

Source: INES2 2016. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

To evaluate whether voters select the party closest to them, we conduct a regression analysis comparing the proximity of a voter to each party with whether they voted for that party. We include two additional control variables. In the first model, we include each party as some parties will, for reasons other than left–right position, receive more support. For example, their relative strength might also be attributed to their historical support levels, likelihood to enter government or other ideological and non-ideological reasons. In the second model, we also add the party’s relative extremism. From the literature on policy discounting (Grofman, 1985), voters may believe a party’s ability to pursue its agenda will be diluted to some degree and so they may counter this by supporting a party with a more extreme position than their ideal policy position (Kedar, 2005). Thus, we include a variable to account for the party’s relative extremism, measured as the party’s absolute deviation from the midpoint of five. The regression analysis reveals a strong relationship between where the respondents located themselves relative to other parties and their choice, suggesting that parties located to within one point of their personal left–right placement were significantly more likely to be chosen as the first preference. Figure 3.3 depicts the relationship between the voter’s party proximity (the absolute difference between the voter’s position and that of a party) and his or her likelihood to support that party. It suggests that almost half of all respondents choose a party within one point of their own position on the left–right scale. The results of this regression analysis are given in Table A3.1 in the Appendix. We also find that voters do tend to vote for parties with more extreme positions than their own, rather than the parties closest to them.

50

The post-crisis Irish voter

Figure 3.3  Effect plots explaining the effect of [1] voter-party proximity and of [2] party extremism on vote choice. Source: INES2 2016.

A dimensional analysis of preference rankings The purpose of this chapter is to investigate the extent to which ideological positioning affects vote choice. One can also reverse the question, and ask whether vote choice can be used to estimate ideological positioning. This is particularly useful as the analysis above does assume that we properly capture the ideological positioning of the respondent with the few attitudinal questions asked in the survey. Perhaps respondents do vote purely for ideological reasons, but their ideology is expressed on different dimensions, or the key dimensions are not accurately measured by the relevant survey questions. Given some assumptions, we can derive a possible ideological positioning of political parties from the relationship between parties in terms of vote transfers. All else being equal, if parties are ideologically close to each other, we expect voters to use their preference votes to support both parties, while if they are ideologically distant, defining major fault lines in political competition, then fewer voters would vote for both parties. We make use not only of the reported first preference vote, but the full preference ordering reported on the ballot in the exit poll.1 This preference ordering is acquired using a mock-ballot paper, which is handed to the respondent at the start of the RTÉ/INES exit poll (INES1) interview. This interview takes place only minutes after the actual voting by the respondent, so that it is likely that the respondent still has a reasonably good recollection of how the real ballot paper was filled out. Table 3.3 gives an overview of the relationship between first and second preference votes. Specifically, we estimate each party’s share of transfers from each other party. We also compare this figure with the

Ideological dimensions in the 2016 elections

51

party’s first preference vote share to see whether a party is attracting more, or fewer, transfers than it might if there were no relationship between parties. We can see, for example, that Sinn Féin attracts 36 per cent of its second preferences from AAA-PBP. If second preferences went according to first preference shares, in terms of the Sinn Féin’s first preference popularity that figure is 2.3 times the number that might be expected. This is indicative of the close relationship between the two parties. In return, AAA-PBP receives a proportion of second preferences from Sinn Féin that is 4.9 times greater than we might expect from the former’s first preference vote share. From our earlier analysis, we know that both parties are most closely defined in terms of left wing, favouring redistribution. Labour and Fine Gael form another grouping, with Fine Gael receiving 2.1 times its vote share in terms of the proportion of second preferences it receives from the Labour Party. In return, the Labour Party receives 4.9 times its first preference share in terms of the share of second preferences from Fine Gael. The two parties stood together in government and are united by a similar level of opposition towards populism. We continue along those lines, but instead of looking at only the first and second preference, we consider the full ballot paper, and use dimensional analysis to provide a visualization of the results. For this analysis we look at the highest preference ranking given by a respondent to each of the parties. For example, if a respondent gives a first preference to a Fine Gael candidate, a second preference to a Labour candidate, a third preference to another Fine Gael candidate and a fourth preference to a Fianna Fáil candidate, then this respondent will score a one for Fine Gael, a two for Labour and a four for Fianna Fáil.2 This leaves, of course, a lot of parties not scored for that particular respondent, since not all parties stand in all constituencies, and most respondents do not rank all candidates on the ballot (for more on this see chapter 2).3 Since the algorithm used only takes ranking into account, not the distance between those rankings, we assign the lowest possible ranking in the dataset – the longest ballot paper has 22 candidates – to all rankings where that particular party did not receive any vote from that particular respondent, whether because it did not stand in the constituency or because the voter decided not to give a vote. Based on these preference scores for each respondent and each party, we can construct a distance matrix between parties. If many respondents give high or low preference rankings to the same pair of parties, those parties have a short distance, and vice versa; when many voters give high preference to one party but low preference to the other, those parties will have a large distance. Such a distance matrix therefore provides a description of the ideological space in which Irish party competition takes place. Parties that are close to each other are competing closely for the same voters, while parties that are further away from each other appeal to very different segments of the voters. To a large extent, we can assume this to be driven by ideological factors – for example, left-wing parties can be

Table 3.4  Party second preference share of party first preferences  

 

Second preference Fine Gael

First preference

  Fine Gael Fianna Fáil Sinn Féin Labour Party AAA-PBP Green Party Social Democrats Independent Other right Other left

-  29 [1.1] 8 [0.3] 54 [2.1] 6 [0.2] 18 [0.7] 22 [0.9] 26 [1.0] 25 [1.0] 7 [0.3]

Fianna Labour Fáil Sinn Féin Party 26 [1.2] 13 [0.6] 16 [0.7] 6 [0.3] 11 [0.5] 15 [0.7] 23 [1.1] 22 [1.0] 6 [0.3]

3 [0.2] 17 [1.1] 4 [0.3] 36 [2.3] 5 [0.4] 13 [0.8] 17 [1.1] 9 [0.6] 16 [1.1]

35 [4.9] 10 [1.4] 4 [0.5] 7 [1.0] 18 [2.6] 10 [1.4] 9 [1.3] 3 [0.4] 12 [1.7]

AAA-PBP

Green Party

Social Democrats

Independent

Other Right

Other Left

2 [0.3] 5 [1.2] 23 [4.9] 3 [0.6] 13 [2.7] 5 [1.0] 6 [1.3] 5 [1.2] 13 [2.8]

5 [1.3] 4 [1.1] 5 [1.5] 6 [1.8] 12 [3.4] 13 [3.7] 4 [1.0] 9 [2.4] 9 [2.5]

4 [1.2] 6 [1.5] 4 [1.1] 5 [1.4] 6 [1.7] 7 [1.8] 4 [1.1] 4 [1.1] 1 [0.4]

19 [1.4] 21 [1.5] 31 [2.3] 10 [0.7] 18 [1.3] 18 [1.3] 14 [1.0] 21 [1.5] 30 [2.2]

5 [1.6] 7 [2.3] 7 [2.1] 1 [0.3] 4 [1.3] 5 [1.5] 5 [1.4] 7 [2.2] 4 [1.4]

2 [1.0] 1 [0.5] 4 [2.1] 1 [0.4] 6 [2.9] 5 [2.5] 3 [1.4] 4 [1.9] 2 [1.2] - 

Percentage of first preference voters for party in the row who give a second preference for party in the column (e.g. 26 per cent of those who voted Fine Gael gave Fianna Fáil as their second preference). In square brackets, the ratio of second preferences compared with the party’s first preference vote share (e.g. The Labour Party’s second preferences from Fine Gael are 4.9 times greater than the Labour Party’s first-preference vote share in this survey). Source: INES1 2016. AAA-PBP = Anti-Austerity Alliance – People Before Profit.

Ideological dimensions in the 2016 elections

53

expected to compete for voters primarily with other left-wing parties. The most expressive technique to interpret such a distance matrix would be to take the spatial analogy further and produce a map, a graphical visualization of the distances. Multidimensional scaling is a statistical technique to reduce such a distance matrix to a two-dimensional representation, such that the relative distances between the different parties are preserved to the greatest extent possible.4 We are thus able to visualize this two-dimensional figure and interpret the visualization as a map of the ideological space in which Irish voting takes place. Figure 3.4 provides this visualization when including the traditional Irish parties and a few that were prominent in the lead-up to the elections, some of which were founded in response to the global financial crisis.5 The first dimension (the dashed diagonal line) clearly aligns with the left–right economicbased ordering of political parties in Ireland. As with a real geographical map, we are free to rotate the map as we see fit, and therefore are free to interpret the diagonal as opposed to the horizontal axis. Indeed, when we look at the self-placement on a left–right scale by the respondents and rank the parties by average self-placement of their (first-preference) voters, we obtain a reasonably similar ordering: from right to left, we would have Fine Gael, Fianna Fáil,

Figure 3.4  Graphical representation of the latent dimensions estimated by the multidimensional scaling algorithm, summarizing distances between parties based on the highest preference for each party assigned by each respondent. Source: INES1 2016.

54

The post-crisis Irish voter

Renua, Social Democrats, Labour, Greens, Sinn Féin and AAA-PBP.6 All the larger parties are classified by their left–right positioning, although smaller parties such the Greens, Renua and the Social Democrats are somewhat to the left of expectations. Results for the smaller parties have to be taken with a grain of salt, however, as the absolute numbers of transfers between these parties among the respondents in the survey are very low indeed. An alternative interpretation of the horizontal axis would be the populist dimension dividing establishment Labour and Fine Gael from anti-establishment Fianna Fáil and Sinn Féin. On the right are the larger, established parties, and on the left, the newer, small parties. Voters may signal an anti-establishment attitude by voting for a number of different parties, ranging from left-wing AAA-PBP to rightwing Renua, simply to indicate dissatisfaction with the establishment parties. This dimensional space is the result of both ideology and other factors that separate voters, including retrospective evaluation of government performance. The second dimension in the figure appears primarily to depict a governmentopposition dimension, with Labour and Fine Gael close to each other and to the party closest to the governing coalition, Fianna Fáil, while Sinn Féin is the most pronounced opposition party. Somewhat related to the governmentopposition dimension is a more ideological dimension based on a populist worldview separating the ordinary people from the elite. As discussed earlier in the chapter, Sinn Féin and Fianna Fáil can be taken as significantly more populist than Labour and Fine Gael, which would be a potential alternative, more ideological explanation of the second axis, in this case perpendicular to the left–right axis in Figure 3.4. To validate our interpretation of the multidimensional scaling results would be to correlate the two resulting dimensions, the two axes of the plot, with demographic and attitudinal variables available on the survey. For the sake of brevity, we leave out the details of the regression analysis.7 The first dimension (the x-axis) correlates weakly with left–right self-placement and more strongly with whether respondents identify as Catholic and their level of church attendance. The second dimension (the y-axis) correlates instead with the left– right self-placement of the respondents. These correlations are particularly weak, however, with various demographics and ideological questions in the surveys only explaining about two per cent or less of the variation in the axes.

A new globalization dimension Given the current attention to the phenomenon of populism and antiglobalization, especially since the election of Donald Trump to the US presidency, the question remains why there is no clear anti-globalization party in Ireland. Elsewhere, this has led to anti-establishment and typically anti-immigration parties on the right, such as the United Kingdom Independent Party, the Front

Ideological dimensions in the 2016 elections

55

Nationale (National Front), the Alternative für Deutschland (Alternative for Germany) or the Partij voor de Vrijheid (Party for the Freedom), or antiglobalization parties on the left, such as Podemos or Syriza. While most of these parties are populist in nature, emphasizing the contrast between policies implemented by the elites and policy preferences of much of the population (Mudde and Kaltwasser, 2017), the ideologically more important commonality is the anti-globalization position, resulting in resistance primarily to liberal immigration policies and to European integration and free trade. Increased international integration is seen as harmful to the local labour market and local culture. De Vries (2017) investigates the relative importance of the economic and cultural components of this cosmopolitan-parochial divide, finding that, at least in the Netherlands, this divide is primarily driven by the relative economic insecurity of individual voters as a consequence of globalization. Given the recent global financial crisis, we could expect Ireland to be fertile ground for such a political party. Some of this electoral space is occupied by Sinn Féin, as quite clearly their supporters do favour some of these policy approaches. While left-wing, it has a clear anti-globalization stance, similar to how the Socialistische Partij (Socialist Party) in the Netherlands is a key competitor to the Partij voor de Vrijheid, or similar to left-wing populist parties such as Podemos and Syriza, although this does not translate into an antiimmigration attitude for Sinn Féin. Indeed, O’Malley (2008) shows that antiimmigration attitudes correlate with increased support for Sinn Féin, despite the pro-immigration stance of the party, suggesting that the party captures to some extent this anti-globalization segment of the population. Whether as a result of a lack of campaigning by Sinn Féin or otherwise, immigration positioning is a very weak predictor of voting behaviour in comparison with other dimensions. The electoral system also affects the party’s ability to manoeuvre. Given the country’s history of emigration and limited immigration, any stance on immigration would almost certainly be divisive. The requirement to win transfers from other parties means that such a divisive issue may damage the party’s electoral prospects. Furthermore, the party must also adopt a coherent stance north and south of the border where it operates. Given that Irish people have for several centuries migrated to Britain, such a position would also be somewhat incoherent. That no other parties or political entrepreneurs emerged is also interesting. It is perhaps in part related to the lack of ideology in the party system in the first place. If the parties are sufficiently pragmatic and populist to adjust to changes in public opinion regarding socio-economic or cultural policies, then there are fewer objectionable policies that anti-establishment parties will react against. Drawing on the regression analysis reported above to determine support for specific parties, we evaluate party support against basic demographics. Sinn Féin is the party of the working class; it is also the party of the less well-educated. Support for Fianna Fáil is also significant in this demographic group. By contrast,

56

The post-crisis Irish voter

Fine Gael and Labour perform significantly better among those with university education and from higher social grades. The groups that are identified by Kriesi et al. (2006) as more vulnerable to international integration and apparently more likely to support more parochial parties are therefore more likely to support the two more nationalist parties, while they are less likely to support Labour or Fine Gael. Using a battery of questions that measure populism more specifically we construct an index based on these questions, and a multinomial regression analysis explaining vote choice confirms this result. Controlling for demographics, left–right position and religiosity, more populist voters are significantly more likely to vote Sinn Féin or Independent, and significantly less likely to vote Fine Gael, than to vote Fianna Fáil. In the dimensional analysis of the full preference ranking on the ballot paper, we also find suggestive evidence of a populist ideological dimension, juxtaposing Sinn Féin and Fianna Fáil to Fine Gael and Labour. Notably, Fine Gael and Labour have indeed significantly different views to Sinn Féin in respect of immigration and one of the key significant differences between Fianna Fáil and Fine Gael relates to immigration.

Conclusion Although ideology may not play the lead in Irish politics, it is at the very least a supporting actor. Crucial to this assertion is the rise of left-wing parties in the aftermath of the financial crisis. In Ireland, the mean electoral support for ‘class left’ parties between the 1950s and 1980s was 13 per cent. This compared with 42 per cent across 14 Western European democracies (Gallagher, Laver and Mair, 1992). In the immediate aftermath of the financial crisis, support for left and centre-left parties has risen to just under 30 per cent in total. This is widely facilitated by support for Sinn Féin – a party which is more closely defined by social class than the Labour Party. Left and right are significant predictors of party support and there appears to be a consistent relationship between the positions adopted by voters and the parties they support. Economic left–right separates Fine Gael supporters from Sinn Féin supporters while socio-moral left–right more clearly divides Fianna Fáil and the Labour Party. Drawing similar conclusions to McElroy’s (2017) analysis of the postcrisis voter, we observe that voters no longer have a difficulty in placing both themselves and political parties on a general left–right dimension. However, unlike McElroy’s analysis that was based on data from 2011, we observe some increasing policy coherence to these placements with voters’ preferences on economic and social issues increasingly corresponding with their own perception of their left and right. While ideological positioning may not separate the two largest parties, it does determine whether someone might vote for either or neither of those parties. On average, Irish voters select parties that are ideologically close to them on a left–right scale. This left–right dimension

Ideological dimensions in the 2016 elections

57

is also clearly dominant in the analysis of the latent ideological space of party competition, where vote transfers are used to estimate the relative ideological distances between parties. When visualizing these distances on a map, we can clearly see the parties ordered along the left–right dimension, in line with expert surveys and self-placement by voters of the parties. Controlling for left and right, those that like Fianna Fáil tend to be more likely in favour of tax and spend policies. The notion that Fianna Fáil is somewhat left of Fine Gael on economic terms should come as little surprise. Fianna Fáil leaders have never shied away from such credentials: De Valera was quoted as saying, ‘In those days I believe we could be called socialists, but not communists’ (Mair, 1987: 17). In the 1980s, Fianna Fáil leader Charles Haughey denounced the 1982–1987 government’s monetarist approach, referring to Keynes as the ‘last economist worth his salt’ (reported in the Irish Times, 6 February 1987). More recently Fianna Fáil Taoiseach Bertie Ahern went so far as to define himself as a ‘socialist’ (Sunday Independent, 31 October 2004). Fianna Fáil appears to be identified by voters as being on the right of the left– right scale on account of its conservative position on socio-moral issues while Fine Gael is more closely associated with economic conservativism. While arguably advocating more profligate and economically left-of-centre policies, Fianna Fáil’s political positioning reflects a populist dimension more than a left-wing economic one. Yet, by contrast to Sinn Féin, the party is still widely considered to be significantly further to the right, economically. Sinn Féin has emerged as a left-wing party. Although there is a significant overlap between socio-moral and economic left and right and globalization, populism offers a clear, second, cross-cutting dimension defining political competition. Sinn Féin and Fianna Fáil supporters are more likely to advocate more populist attitudes whereas Fine Gael and Labour supporters are more likely to advocate more elitist attitudes. Although immigration and globalization do not currently play a significant role in Irish politics, they map onto the Irish political system as somewhere between left–right and populist dimensions. One significant caveat is that this analysis reflects a period in which Fine Gael and Labour were in government. Thus, the most clearly identifiable distinction in the latent space analysis is more simply, and more idiosyncratically, the contrast between the government of the day and the opposition. The Irish case in a comparative perspective may appear more conventional in terms of left–right competition than typically assumed; it also has an undercurrent of anti-globalization similar to that found in other European states. The flexible nature of the two catch-all parties that dominate the party system may lead to further development of this dimension. However, the stability of Irish politics suggests that movement in this direction may very well be slow. That, of course, depends on whether Irish politics has entered a period of continued instability or whether it is in a shorter phase of re-alignment.

Chapter Appendix Table A3.1  Modelling the relationship between a respondent’s party choice and the difference between where they put themselves and where they put each party on the left–right scale Fixed effects: Intercept Fianna Fáil Fine Gael Green Labour Sinn Féin Social Democrats Party-self left–right difference Party distance from centre Source: INES2 2016. SE = standard error.

Est. −2.300 1.988 2.573 0.467 1.009 1.747 0.416 −0.481

SE

P-Value

0.248 0.260 0.258 0.287 0.275 0.268 0.295 0.031

0.000 0.000 0.000 0.104 0.000 0.000 0.158 0.000

Est. *** *** *** *** *** ***

−2.868 2.171 2.705 0.786 1.350 1.940 0.660 −0.555 0.244

SE

P-Value

0.264 0.264 0.261 0.293 0.281 0.272 0.299 0.033 0.034

0.000 0.000 0.000 0.007 0.000 0.000 0.027 0.000 0.000

*** *** *** ** *** *** * *** ***

Ideological dimensions in the 2016 elections

59

Notes 1 In this manner, we make use of the unique feature of STV electoral systems whereby we have relative preferences for a range of parties for each voter to perform our ideological analysis of voting behaviour. An alternative, for example, would be to use sympathy scores, measures of how much a voter likes each party or probability-to-vote questions. While these are perfectly valid alternatives, an advantage of the approach presented here is that generally in surveys, questions about behaviour – such as how a ballot was filled out – tend to be somewhat more reliable than attitude questions or hypothetical questions such as the probability to vote. 2 One weakness of this approach is that voters will base their preference on not only the party, but also the candidate (see chapter 10), which intervenes with the ideological interpretation of the party ordering. 3 Our measure thus provides a measure of the highest preference given to each party, based on the candidates ranked. An alternative approach would be to simply rank the parties, ignoring intra-party transfers. For example, in our analysis, a voter who ranks candidates from, respectively, Fine Gael, Labour, Fine Gael and Fianna Fáil would in our data obtain a lower score for Fianna Fáil than someone who ranks them as Fine Gael, Labour, Fianna Fáil and Fine Gael, while if simply parties were ranked, scores would be identical. We argue that the former voter, if ideology drives the ranking, would consider Fianna Fáil ideologically further away than the latter voter does. 4 We make use of the MDPREF algorithm which allows for multidimensional scaling with preference data, using the implementation in the ‘pmr’ package in R by Paul H. Lee and Philip L. H. Yu. 5 When all parties are included, all the very small parties end up in the same part of the ideological space as Renua and the Social Democrats, except for the Independent Alliance, which ends up somewhere between Sinn Féin and Fianna Fáil. 6 The rank ordering based on the 2014 data of the Chapel Hill Expert Survey also confirms this left–right ordering of Irish political parties. http://chesdata.eu/, accessed 1 August 2017. 7 Available upon request from the authors.

References Benoit, Kenneth, and Michael Laver. 2003. ‘Estimating Irish party policy positions using computer wordscoring: The 2002 election–a research note’, Irish Political Studies. 18: 97–107. Bowler, Shaun, and David M. Farrell. 2017. ‘The lack of party system change in Ireland in 2011’, in Michael Marsh, David Farrell and Gail McElroy (eds), A Conservative Revolution? Electoral Change in Twenty-First Century Ireland. Oxford: Oxford University Press, pp. 83–101. Breen, Richard, and Christopher T. Whelan. 1994. ‘Social class, class origins and political partisanship in the Republic of Ireland’, European Journal of Political Research. 26: 117–33. Carty, R. Kenneth. 1981. Electoral Politics in Ireland: Party and Parish Pump. Ontario: Wilfrid Laurier University Press. Chubb, Basil. 1957. ‘The independent member in Ireland’, Political Studies. 5: 131–39.

60

The post-crisis Irish voter

De Vries, Catherine E. 2017. ‘Benchmarking Brexit: How the British decision to leave shapes EU public opinion’, Journal of Common Market Studies. 55: 38–53. Farrell, David, 1999. ‘Ireland: A party system transformed?’, in David Broughton and Mark Donovan (eds), Changing Party Systems in Western Europe. London: Pinter, pp. 30–47. Gallagher, Michael. 1985. Political Parties in the Republic of Ireland. Manchester: Manchester University Press. Gallagher, Michael, and Lee Komito. 2010. ‘The constituency role of Dáil deputies’, in John Coakley and Michael Gallagher (eds), Politics in the Republic of Ireland. London: Routledge, pp. 230–62. Gallagher, Michael, Michael Laver and Peter Mair. 1992. Representative Government in Western Europe. New York: McGraw Hill. Gallagher, Michael, and Michael Marsh. 2002. Days of Blue Loyalty: The Politics of Membership of the Fine Gael Party. Dublin: PSAI Press. Goodhart, David. 2017. The Road to Somewhere: The Populist Revolt and the Future of Politics. Oxford: Oxford University Press. Grofman, Bernard. 1985. ‘The neglected role of the status quo in models of issue voting’, Journal of Politics. 47(1): 230–37. Jennings, Will, and Gerry Stoker. 2017. ‘Tilting towards the cosmopolitan axis? Political change in England and the 2017 General Election’, Political Quarterly. 88(3): 359–69. Kedar, Orit. 2005. ‘When moderate voters prefer extreme parties: Policy balancing in parliamentary elections’, American Political Science Review. 99(2): 185–99. Kriesi, Hanspeter, Edgar Grande, Romain Lachat, Martin Dolezal, Simon Bornschier and Timotheos Frey. 2006. ‘Globalisation and the transformation of the national political space: Six European countries compared’, European Journal of Political Research. 45: 921–56. Kriesi, Hanspeter, Edgar Grande, Romain Lachat, Martin Dolezal, Simon Bornschier and Timotheos Frey. 2008. West European Politics in the Age of Globalisation. Cambridge: Cambridge University Press. Laver, Michael. 1992. ‘Are Irish parties peculiar?’, in John Goldthorpe and Christopher Whelan (eds), The Development of Industrial Society in Ireland. Oxford: Oxford University Press, pp. 359–81. Laver, Michael, Peter Mair and Richard Sinnott (eds). 1987. How Ireland Voted: The Irish General Election, 1987. Dufour Editions, Dublin: Poolbeg. Lipset, Seymour M., and Stein Rokkan. 1967. ‘Cleavage structures, party systems, and voter alignments: An introduction’, in Seymour M. Lipset and Stein Rokkan (eds), Party Systems and Voter Alignments: Cross-National Perspectives. New York: Free Press, pp. 1–64.

Ideological dimensions in the 2016 elections

61

Lutz, Karin G. 2003. ‘Irish party competition in the new millennium: Change or plus ça change?’, Irish Political Studies. 18(2): 40–59. Mair, Peter. 1986. ‘Locating Irish political parties on a left-right dimension: An empirical enquiry’, Political Studies. 34(3): 456–65. Mair, Peter. 1987. The Changing Irish Party System: Organisation, Ideology and Electoral Competition. London: Pinter. Mair, Peter. 2011. ‘The election in context’, in Michael Gallagher and Michael Marsh (eds), How Ireland Voted 2011. London: Palgrave Macmillan, pp. 283–97. Marsh, Michael. 2006. ‘Party identification in Ireland: An insecure anchor for a floating party system’, Electoral Studies. 25(3): 489–508. Marsh, Michael. 2007. ‘Candidates or parties? Objects of electoral choice in Ireland’, Party Politics. 13(4): 500–27. Marsh, Michael, David M. Farrell and Gail McElroy (eds). 2017. A Conservative Revolution? Electoral Change in Twenty-First Century Ireland. Oxford: Oxford University Press. Marsh, Michael, and Kevin Cunningham. 2011. ‘A positive choice, or anyone but Fianna Fáil?’, in Michael Gallagher and Michael Marsh (eds), How Ireland Voted 2011. London: Palgrave Macmillan, pp. 172–204. Marsh, Michael, Richard Sinnott, John Garry and Fiachra Kennedy (eds). 2008. The Irish Voter: The Nature of Electoral Competition in the Republic of Ireland. Manchester: Manchester University Press. Mayda, Anna Maria, and Dani Rodrik. 2005. ‘Why are some people (and countries) more protectionist than others?’, European Economic Review. 49(6): 1393–1430. McElroy, Gail. 2017. ‘Party competition in Ireland’, in Michael Marsh, Richard Sinnott, John Garry and Fiachra Kennedy (eds), 2008. The Irish Voter: The Nature of Electoral Competition in the Republic of Ireland. Manchester: Manchester University Press, pp. 61–82. Meyer, David S., and Sidney Tarrow. 1998. ‘A movement society: Contentious politics for a new century’, in John D. McCarthy, Clark McPhail, David S. Meyer, and Sidney Tarrow (eds), The Social Movement Society: Contentious Politics for a New Century. London: Rowman and Littlefield, pp. 1–28. Mudde, Cas. 2004. ‘The populist zeitgeist’, Government and Opposition. 39(4): 542–63. Mudde, Cas, and Cristóbal Rovira Kaltwasser. 2017. Populism: A Very Short Introduction. Oxford: Oxford University Press. O’Malley, Eoin. 2008. ‘Why is there no radical right party in Ireland?’, West European Politics. 31(5): 960–77. O’Rourke, Kevin H., and Richard Sinnott. 2006. ‘The determinants of individual attitudes towards immigration’, European Journal of Political Economy. 22(4): 838–61.

62

The post-crisis Irish voter

Oppenhuis, Erik. 1995. Voting Behavior in Europe: A Comparative Analysis of Electoral Participation and Party Choice. Amsterdam: Het Spinhuis. Sinnott, Richard. 1995. Irish Voters Decide: Voting Behaviour in Elections and Referendums since 1918. Manchester: Manchester University Press. Teney, Céline, Onawa Promise Lacewell and Pieter de Wilde. 2014. ‘Winners and losers of globalization in Europe: Attitudes and ideologies’, European Political Science Review. 6(4): 575–95. Urwin, Derek W., and Kjell A. Eliassen. 1975. ‘In search of a continent: The quest of comparative European politics’, European Journal of Political Research. 3(1): 85–113. Van der Eijk, Cees, and Mark N. Franklin. 1996. Choosing Europe? The European Electorate and National Politics in the Face of Union. Ann Arbor, MI: University of Michigan Press. Whyte, John. 1974. ‘Ireland: Politics without social bases’, in Richard Rose, Electoral Behaviour: A Comparative Handbook. New York: The Free Press, pp. 619–51.

4 Social and ideological bases of voting John Garry

Introduction In contrast to our near neighbours – Northern Ireland and Britain – Ireland is often characterized as having a political system in which voting behaviour is only loosely connected to the socio-demographic and ideological traits of voters. In the British case, one academic (Pulzer, 1967: 98) famously stated that ‘class is the basis of British party politics; all else is embellishment and detail’. This characterization still resonates today, with the working class, who tend to have economically left-wing beliefs, voting for Labour while the middle class, who are more likely to hold relatively pro-free-market views, tend to support the Conservative party (Evans and Tilley, 2017). In Northern Ireland, the main ‘political cleavage’ is even more potent. Citizens from a Catholic community background are much more likely than citizens from a Protestant community background to hold pro-United Ireland views and vote for nationalist parties, while pro-UK Protestants overwhelmingly support unionist parties (Garry, 2016). In Ireland, there is no analogous political cleavage that allows us to easily make sense of how Irish people vote. The choice of which party to vote for at an Irish election is much less anchored in social traits and ideological beliefs linked to those traits (Marsh et al., 2008). In fact, for much of Ireland’s history, the burning question for observers at election time has been: Is there any difference between the two main parties, Fianna Fáil and Fine Gael? The initial ‘civil war’ distinction between the parties, essentially pitting the strongly nationalist Fianna Fáil against the moderately nationalist Fine Gael, has long since faded. Furthermore, the dominance of this ‘national question’ at the time of the formation of the state also served to constrain the emergence of a clear economic left versus right distinction in the party system, resulting in the Labour party being unable to make strong headway as a party of the working class seeking to implement leftist beliefs (Sinnott, 1984). Whyte’s much-quoted characterization

64

The post-crisis Irish voter

of Ireland in the early 1970s as ‘politics without social bases’ still has significant truth (Whyte, 1974). Many analysts have examined the empirical basis of Irish voting and have found at least ‘some’ structure to the vote (for example, Laver’s ‘Politics with some social bases’ (1986); see also Marsh and Sinnott, 1990; 1993; 1999; Sinnott, 1995; Garry et al., 2003; Marsh, 2007; Marsh and Cunningham, 2011). However, its limited nature is perhaps best exemplified by the finding of the first Irish National Election Study, that Labour tended to attract support more from the middle than the working class (Marsh et al., 2008). In a recent study of the 2007–2011 election time points, Tilley and Garry (2017) did find significant evidence of a class structure to the vote, suggesting that Ireland’s economic crisis, and the possible policy responses to it, may have jolted the political system towards the international norm of a choice between interventionist parties (Labour and Sinn Féin) and free-market parties (Fine Gael), leaving a defeated catch-all (Fianna Fáil) behind in the electoral rubble. Tilley and Garry found that the vote choice of Fianna Fáil defectors could be explained in class and economic policy terms: relatively middle-class and rightwing votes went to Fine Gael while the more working-class and leftist votes went to Labour and Sinn Féin. As the authors acknowledge, this class structure to the vote may have been due to the unique circumstances of that election, in which a ‘catch-all’ centrist party imploded due to widely perceived mismanagement of the economy. One of the aims of this chapter is to continue this examination of the role played by class and economic policy beliefs and to do so in the postcrisis context of the 2016 election. Aside from economic matters, one cluster of issues in Ireland that has proved to be socially and ideologically rooted, and also highly salient politically, relates to the ‘liberal versus conservative’ dimension in politics – including issues such as abortion and gay rights. Highly religious voters tend to hold socially traditional views while non-religious or minimally religious voters are of a more liberal disposition. How this distinction links to voting, however, has typically been evident in the context of referendums rather than elections, with several referendums since the early 1980s on abortion and also on divorce and marriage equality, and another referendum on abortion likely in 2018. Political parties have indeed adopted positions on these issues and insofar as the liberal-conservative dimension does influence election campaigns and party competition, a relatively liberal Labour party has competed against the more conservative-minded Fianna Fáil and Fine Gael parties (Marsh et al., 2008). In order to assess whether this distinction between Labour and the two main parties still holds, the role played by socio-moral issues in structuring vote choice in 2016 will be examined in this chapter. While these three parties – Labour, Fianna Fáil and Fine Gael – have typically dominated Irish elections, Sinn Féin can now be added as a ‘major player’, achieving a significant increase in vote share and seat share in 2016

Social and ideological bases of voting

65

and replacing Labour as the third biggest party in the state. This – seemingly sustainable – rise of Sinn Féin suggests an increased role played by social and ideological factors in the structuring of Irish voting, as the party has clearly defined positions on all three policy themes of economy, nation and sociomoral issues. Most obviously, Sinn Féin is likely to attract votes from strongly nationalist citizens, but also – given its economic leftism and social liberalism – from left-wing/working-class citizens and from those on the socially liberal side of the socio-moral debate (see discussion in Garry, 2017). The overall aim of this chapter is to relate these broad ideological themes – economy, nation and socio-moral values – to vote choice at the 2016 election in order to paint a picture of the post-crisis Irish voter. Is there a strong link between voters’ socio-demographic traits, their broad policy beliefs and their party choice? We begin by briefly outlining some expectations regarding the social and ideological bases of party support. We then describe the data that we will use in our empirical investigation of these expectations. We report the social and ideological characteristics of the supporters of the different parties in 2016. We then focus on change over time and assess the social and ideological profile of voters who switched from the main governing party (Fine Gael) in 2016 to one or other of the rival main parties.

Expectations A simple ‘model’ of voting is sketched in Figure 4.1, in which choice at election time is driven by one’s everyday social experience and behaviour, and the political beliefs that are likely to flow from that experience and behaviour. On the theme of the economy, citizens are expected to vary in terms of the extent to which they are either on the free market or socialist side of the ‘left versus right’ debate. Citizens with leftist views, who favour a strong interventionist state to reduce inequality and fund state services, are likely to come from a working-class background, while citizens who are likely to benefit most from a more laissez-faire economic approach are likely to be middle class. Insofar as these factors influence voting, we expect the biggest differences to be between the leftist Sinn Féin and the much more economically conservative Fine Gael. On the theme of morality, a socially conservative position is likely to be held by strongly religious citizens while social liberals are likely to come from non-religious backgrounds. We expect the traditionally conservative Fianna Fáil to be most distinct from the liberal Labour party. We suggest that both religion and education are likely to undergird beliefs on the national question, with Catholics1 and the less educated being more likely to hold nationalist positions and, in turn, support Sinn Féin as the most clearly nationalist party. In contrast, being non-Catholic and highly educated is likely to dispose one

66

The post-crisis Irish voter vote choice

Theme

Social base

ideology

Economy

social class

free market vs. socialist

Middle class

free market

Fine Gael

Working class

socialist

Sinn Féin

religiosity

Liberal vs. conservative

Highly religious

conservative

Fianna Fáil / Fine Gael

Not religious

liberal

Labour

religion + education

strongly vs. weakly nationalist

Catholic

strongly nationalist

Sinn Féin

Not Catholic

weakly nationalist

Fine Gael

Low educational level

nationalist

Sinn Féin

High educational level

unionist

Fine Gael

Morality

Nation

Figure 4.1  Expectations of the social and ideological bases of vote choice. Note: Only the strongest expectations are sketched here.

to eschew a nationalist position and hence being more inclined to support the more pluralist Fine Gael. This model is relatively sparse as it focuses on only a small number of sociodemographic factors and hypothesizes only the strongest expected relationships regarding the social and ideological underpinnings of the vote. We additionally seek to examine the importance of factors such as age, gender and place of residence, and we also examine the Greens, AAA-PBP and Independents, but Figure 4.1 acts as a valuable heuristic guide to frame our examination of the key social and ideological expectations regarding the main parties in the system.

Measurement All of the survey data analysed in this chapter come from one survey, namely the RTÉ/INES exit poll (INES1). The wording of the survey questions that were designed to tap citizens’ stances on the three main ideological themes is reported

Social and ideological bases of voting

67

in Figure 4.2. Two survey items are used on the economy: attitudes to taxation and spending and attitudes to equality. Abortion is used as the key issue to measure positions on the liberal–conservative continuum, and a question about the desired long-term constitutional position of Northern Ireland is asked to measure citizens’ dispositions on the nationalist–unionist debate. In order to enable a simple analysis, we recode the responses of each of these four survey questions into a smaller number of categories (as detailed in Figure 4.2). Ideally, in order to generate a high-quality measure of citizens’ attitudes on a particular ideological dimension, a range of survey questions would be asked and a summary measure would be generated from this set of questions (see, for example, Evans et al. (1996) on multi-item scales measuring economic left–right and libertarian–authoritarian values). Having only

Socio-Moral On a scale from 0 to 10 where 0 means you strongly believe that there should be a total ban on abortion in Ireland, and 10 means that you strongly believe that Abortion should be freely available in Ireland to any woman who wants to have one, where would you place your view? Nation In terms of the long term future of Northern Ireland, which would you prefer? Northern Ireland should: Remain in the UK with a direct and strong link to Britain Remain in the UK and have a strong Assembly and Government in Northern Ireland Unify with the Republic of Ireland Economy Taxation and spending On a scale of 0 to 10, where ‘0’ means government should CUT TAXES A LOT and SPEND MUCH LESS on health and social services, and ‘10’ means government should INCREASE TAXES A LOT and SPEND MUCH MORE on health and social services. Where would you place yourself in terms of this scale? Equality On a scale from 0 to 10 where 0 means you strongly believe that the government SHOULD ACT to reduce differences in income and wealth, and 10 means that you strongly believe that the government SHOULD NOT ACT to reduce differences in income and wealth, where would you place your view? Note – For simplicity of analysis these variables are recoded. Socio-moral: 03=conservative, 5-6=centrist, 7-10=liberal. Nation: 1-2=UK, 3=Unity. Taxation and Spending: 0-3=cut taxation and spending, 4-6=centrist, 7-10=increase taxation and spending. Equality: 0-3=reduce inequality, 4-6=centrist, 7-10=not reduce inequality.

Figure 4.2  Wording of the ideology questions.

68

The post-crisis Irish voter

one question on the liberal-conservative theme and one question on the nationalism theme may be seen as somewhat crude in measurement terms. However, these particular single items were chosen to capture a core element of the theme that is of significant political importance. On the economic theme, we capture two distinct components of the general social democratic versus free-market approach to economic management, and these focus on equality and taxation and spending. For consistency, these are used as single item measures. A different approach to assessing ideology and voting is to ask voters how they position themselves on a ‘left’ versus ‘right’ scale. This has the attractiveness of providing a single item overall summary of citizens’ policy views. However, what we are keenly interested in examining in this chapter is the relative strength of distinct ideological dimensions (economic, nation and social/ moral) in explaining vote choice. While useful in many ways, it is not possible to disentangle the distinct effects of the different ideological components of a single ‘left’ versus ‘right’ measure (see Cunningham and Elkink in chapter 3 of this volume; see also Inglehart and Klingemann, 1976; Knutsen, 1997). Furthermore, as McElroy (2017) has shown using a number of different surveys, only to a weak extent does left–right self-placement act as a proxy for policy positions for voters in Ireland.

Ideological profile of Irish voters As illustrated in Figure 4.3, just over one third of voters favour Irish unity while almost half prefer that Northern Ireland remains in the United Kingdom (either in the form of direct rule or in the context of a functioning devolved administration). A quite high proportion of respondents do not have a clear view either way (17 per cent). Voters’ responses to the question on abortion availability suggest a much greater proportion of liberals than conservatives in Ireland on this issue, with one fifth on the conservative side of the debate and almost one half on the liberal side. Irish voters emerge as quite economically left wing in terms of their responses to questions about tax and spend and equality. Over a third favour greater taxation and spending compared with only one in ten who adopts a right-wing anti-tax approach. Almost two fifths favour government action to reduce inequality compared with less than a quarter who are opposed to such government intervention. Overall, the picture emerges of Irish voters as weakly nationalist, quite strongly liberal and more likely to indicate support for interventionist than free-market economic policy options.

69

Social and ideological bases of voting 60

Northern Ireland future?

50

50

40

40

30

30

20

20

10

10

0

Unity

UK

other/dk

Reduce Inequality?

0

conservative

centrist

liberal

dk

Taxation and spending?

60

60

50

50

40

40 30

30

20

20

10

10

0

Abortion availability?

60

reduce inequality

centrist not reduce inequality

dk

0

cut tax+spend

centrist more tax+spend

dk

Figure 4.3  Voters’ ideological beliefs. Source: INES1 2016.

Social bases of ideological beliefs Is there a clear social basis to these ideological beliefs? Most strikingly, in relation to economic beliefs, the answer is ‘no’. As reported in Table 4.1, there are no significant social-class differences between those who favour left-wing positions on tax and equality and those who do not: 36 per cent of the middle class (ABC1) favour more tax and spend, and an almost identical figure of 37 per cent of the working class (C2DE) favour more tax and spend. The percentage difference is slightly larger for the issue of equality, but it is not statistically significant. In fact, the only social basis that emerges for the economic questions suggests that older voters and urban voters hold relatively left-wing views on taxation and spending. The picture emerging, then, is of variation among voters on economic left–right ideological issues, but the views voters hold are not rooted in, or explained by, their socio-demographic profile. In terms of a possible social basis to views on the national question, our expectations are consistent with the data. Support for a united Ireland is less likely to come from the highly educated and is more likely to come from those with lower levels of education, and it is also more likely to come from Catholics than non-Catholics. In fact, the percentage support for unity among

70

The post-crisis Irish voter

Table 4.1  Social bases of ideological beliefs Pro-united Ireland %

Socially liberal %

More tax and Reduce spend % Inequality %

All Age 18–44 45+

37 35 38

46 54 41

36 32 40

38 38 38

Female Male

31 42

48 44

36 37

38 38

ABC1 C2DE F

34 40 32

52 41 21

36 37 26

35 40 44

Third-level Not third-level

32 40

54 40

38 35

35 40

Catholic Not Catholic

38 31

41 64

36 40

39 34

Weekly or more Less/never

35 37

28 56

37 36

36 39

Rural Urban

38 36

35 52

29 40

35 39

Source: INES1 2016. Note: Figures relate to the percentage of each opinion group made up of voters with the associated socio-demographic trait. For example, 35 per cent of 18–44-year-olds favour a united Ireland and 38 per cent of 45+ year-olds favour a united Ireland. Italicized differences are statistically significant at 0.05 level or better (e.g. male respondents are statistically significantly more likely to favour a united Ireland than females [42 per cent of males compared with only 31 per cent of females]). ABC1 = middle-class demographic; C2DE = working-class demographic; F = farmers.

low-attending Catholics (39.6 per cent) is higher than that among frequently attending Catholics (35.2 per cent), suggesting that it is the group identification as Catholic that matters rather than intensity of Catholic religious belief and engagement. A much stronger social basis for views on the liberal-conservative dimension is evident. In line with expectations, voters who regularly attend religious services (weekly or more frequently) are much less likely to be socially liberal (28 per cent) than voters who are infrequent attenders or who never attend (56 per cent). Furthermore, strong differences emerge on almost all of our social factors: liberals are likely to be younger, middle class, third-level educated, non-Catholic and urban.

71

Social and ideological bases of voting

In order to further tease out which particular socio-demographic characteristics are the most powerful predictors of ideological beliefs, a series of analyses were conducted which, for each ideological theme, used all the social factors simultaneously in a single model (see Table 4.2).2 It emerges that frequent attendance at religious services is the most potent explanatory factor, while Table 4.2  Social bases of ideological beliefs: Binary logistic regression models Pro-united Ireland Age 18–44 Female ABC1 F Third-level Catholic Weekly or more Rural Constant N Nagelkerke r-square

0.92 0.65*** 0.92 0.67 0.74* 1.36* 0.81 1.1 0.72* 1492 0.03

Socially liberal

More tax and spend

1.16 1.24 1.24 0.66 1.23 0.58*** 0.41*** 0.65*** 1.41* 1492 0.14

0.68** 0.97 0.89 0.68 1.19 0.82 1.11 0.65*** 0.87 1492 0.03

Reduce inequality 1.02 0.99 0.83 1.37 0.87 1.32 0.81 0.80 0.65** 1492 0.01

Source: INES1 2016. Note: Figures reported are odds ratios (*0.05 level, **0.01 level, ***0.001 level). ABC1 = middle-class demographic; F = farmers. Explanatory power of social factors

% variation explained

16 14 12 10 8 6 4 2 0

unity

liberal

more tax/spend

reduce inequality

Figure 4.4  Extent to which ideological views can be explained by sociodemographic characteristics. Source: INES1 2016.

72

The post-crisis Irish voter

being Catholic and living in a rural area are also strong predictors of holding conservative beliefs. As illustrated in Figure 4.4, the explanatory power of this model is quite strong, explaining 14 per cent of the variation, and importantly dwarfing the ability of social factors to explain either the nation or economy factors (which have associated explanatory power of between one and three per cent). However, while our two expected predictors of being pro-unity – being Catholic and having a low education – do emerge as significant, and consistent with Table 4.1, no class basis to economic left–right beliefs emerges, suggesting economic attitudes that are free-floating and removed from the class basis typically found in other democracies (see discussion in Evans and Tilley, 2017).

Social bases of voting We now focus on vote choice and examine how these social and ideological variables are systematically related to opting for one particular party rather than another. We begin by providing a simple socio-demographic profiling of the type of people who vote for each party (see Figure 4.5). In terms of the age profile, what is striking is that the three ‘traditional’ parties of Fianna Fáil, Fine Gael and Labour have a lower proportion of young voters than the more youth-friendly Sinn Féin, the Greens and AAA-PBP – with the Independents in between (see also chapter 5 in this volume on the Sinn Féin age profile). The social-class profile of each party’s voters suggests that only about one third of Sinn Féin voters are middle class compared with almost 60 per cent of Fine Gael and Labour voters, and over 70 per cent of Green voters. AAA-PBP and Fianna Fáil share a similar proportion of middleclass voters (approx. 47 per cent), while Independents tend to have less middleclass support. The pattern of the relationship is almost identical when we look at voters with high education levels. When we examine the Catholic and religiosity variables, both graphs (Figure 4.6a and b) show the same relationship, with the exception of Sinn Féin support. Just over four fifths of Sinn Féin voters are Catholic and this is the same proportion as for Fine Gael and Independents and only a few percentage points lower than for Fianna Fáil. Labour is significantly lower and the Greens are particularly low. But in terms of religious attendance, only one fifth of Sinn Féin voters regularly attend and the relative position of the other parties is approximately the same as the Catholic distribution. A significant proportion of Sinn Féin voters, then, are non-religious Catholics. This is consistent with the idea that Catholic identity (rather than belief or behaviour) is a driver of Sinn Féin support. In terms of place of residence, Fine Gael, Fianna Fáil and Independents have over 36 per cent of supporters who are rural dwellers, compared with only seven per cent of AAA-PBP voters.

73

Social and ideological bases of voting % of each party’s voters who are aged 18–44 80 70 60 50 40 30 20

FF

FG

Lab

GR

SF

AAA/PBP Non-party

% of each party’s voters who are middle class

80 70 60 50 40 30 20 FF

FG

Lab

GR

SF

AAA/PBP Non-party

% of each party’s voters with third level education

80 70 60 50 40 30 20

FF

FG

Lab

GR

SF

Figure 4.5  Party voters by age, class and education. Source: INES1 2016.

AAA/PBP Non-party

74

The post-crisis Irish voter

% of each party’s voters who are Catholic

100 90 80 70 60 50 40 30

FF

FG

Lab

GR

SF

AAA/PBP Non-party

% of each party’s supporters who attend weekly or more

60 50 40 30 20 10 0

FF

FG

Lab

GR

SF

AAA/PBP Non-party

% of party’s voters who live in rural areas

45 40 35 30 25 20 15 10 5 0

FF

FG

Lab

GR

SF

AAA/PBP Non-party

Figure 4.6  Party voters by Catholic, religiosity and residence. Source: INES1 2016.

Social and ideological bases of voting

75

Ideological bases of voting There are also some notable differences in the ideological profiles of each party’s voters. Between 57 and 64 per cent of Labour, Green and AAA-PBP voters are socially liberal compared with 43 per cent or less of Fine Gael, Fianna Fáil and Independent voters. More sizeable differences are evident when one looks at the percentage of each party’s voters who support a united Ireland: almost 60 per cent of Sinn Féin voters compared with between 27 and 33 per cent of Fianna Fáil, Fine Gael and Labour voters (and pro-unity Green voters are particularly rare). The pattern emerging from the two economic ideology factors is inconsistent: while Labour, the Greens and AAA-PBP voters emerge as the most left wing on taxation and spending, it is AAA-PBP and Sinn Féin whose voters tend to be left wing on the issue of equality. Taking both demographics and ideology into account, it is worth reflecting on the extent to which the profile of the parties divides neatly into two blocks: the ‘old’ or traditional parties of Labour, Fine Gael and Fianna Fáil on the one hand and the ‘new’ or ‘anti-establishment’ parties on the other (Sinn Féin, AAAPBP, Greens and Independents). Something approaching a pattern emerges in relation to age (the traditional parties have an older age profile) and religiosity (the traditional parties have a church-attending profile). Yet Sinn Féin looks more like an older party in terms of having a strongly Catholic profile and being quite rural. This possible ‘two-block’ interpretation is also hampered by the differences between the Greens and AAA-PBP (on education and class for example) and differences between Labour on the one hand and Fine Gael/ Fianna Fáil on the other (on extent of liberal attitudes for example).

Social and ideological model of voting In order to be more precise about the key socio-demographic and ideological predictors of vote choice, we enter the full set of variables into a series of regression models – reported in Table 4.3 – and relate our findings directly to our expectations set out in Figure 4.1. We begin by comparing Fine Gael voters with Sinn Féin voters. In line with our expectation that this comparison would be the most distinct with respect to social and ideological predictors, we find the model explains almost 40 per cent of variation (joint most predictive of the six models). In line with expectations, we find a class, education and religious attendance effect. The class effect is the strongest, with working-class voters being much more likely to support Sinn Féin than Fine Gael, and also those with third-level education and frequent religious attendance more likely to support Fine Gael than Sinn Féin. A very strong pro-unity effect emerges, with those favouring a united Ireland being much more likely to support Sinn Féin than Fine Gael, and there is also an effect of holding left-wing views on

% of each party’s voters who are socially liberal

70 65

60

60 55

50

50

40

45 40

30

35

20

30

10

25 20

FF

FG

Lab

GR

SF

AAA/PBP Non-party

% of each party’s voters who favour more tax+spend

55

0

60

45

50

40

40

35

30

30

20

25

10 0 FF

FG

Lab

GR

SF

AAA/PBP Non-party

Figure 4.7  The ideological profile of each party’s voters. Source: INES1 2016.

FF

FG

Lab

GR

SF

AAA/PBP Non-party

% of each party’s voters who favour reducing inequality

70

50

20

% of each party’s supporters who favour Irish unity

70

FF

FG

Lab

GR

SF

AAA/PBP Non-party

77

Social and ideological bases of voting

equality relating to voting for Sinn Féin rather than Fine Gael. Overall, this may be regarded as quite a strong model of the social and ideological influence on vote choice between these two parties. An almost identical, albeit not quite as strong, model separates Sinn Féin voters from Fianna Fáil voters (the main difference being the absence of a thirdlevel effect and an equality effect). What is interesting, when one compares Sinn Féin voters to Labour voters, is that Sinn Féin voters are considerably more likely than Labour voters to be working class, and are more likely to be Catholic, but less likely to frequently attend religious services. Sinn Féin voters are also, unsurprisingly, much more likely than Labour voters to be pro-unity, but are also considerably more likely than Labour voters to favour a reduction in inequality. Thus, in the comparisons between Sinn Féin and the three other main parties, quite significant social and ideological bases to voting emerge. When the three other parties – Fine Gael, Fianna Fáil and Labour – are each compared with each other, the explanatory power of the social and ideological factors is much less (between 8 and 18 per cent). The most noteworthy model is the Table 4.3  Social and ideological bases of vote choice: Binary logistic regression models

Age 18–44 Female ABC1 F Third-level Catholic Weekly or more Rural Unity Liberal More tax and spend Reduce inequality Constant N Nagelkerke r-square

SF not FG

SF not FF

SF not Lab

FF not FG

Lab not FG

Lab not FF

2.36*** 0.87 0.23*** 0.20** 0.46** 1.05 0.47** 0.44*** 3.57*** 1.18 1.08 1.63* 1.01 596 0.38

1.99*** 1.16 0.45*** 0.66 0.64 0.90 0.30*** 0.50*** 2.69*** 1.09 1.06 1.27 0.99 549 0.28

2.40*** 0.95 0.18*** 1.32 0.53 2.81** 0.29*** 0.77 3.73*** 0.85 0.78 2.27** 1.95 334 0.38

1.08 0.80 0.60** 0.35** 0.70 1.56 1.35 0.98 1.22 0.96 0.96 1.33 0.85 669 0.08

0.90 0.78 1.11 0.13 0.77 0.43** 1.25 0.56* 1.27 1.63 1.50 0.95 0.47 454 0.11

0.88 0.97 2.07* 0.39 1.17 0.31*** 1.01 0.60 0.86 1.66 1.27 0.70 0.49 407 0.18

Source: INES1 2016. ABC1 = middle-class demographic; F = farmers.

78

The post-crisis Irish voter

comparison between Fianna Fáil and Labour, where Labour’s middle class and non-Catholic credentials shine through. Fianna Fáil also emerges as being significantly more working class than Fine Gael.

Explaining voter dynamics: Vote choice of Fine Gael deserters Another way to investigate the possible social and ideological bases of vote choice is to focus on over-time dynamics. Following the style of analysis of Tilley and Garry (2017), we hone in on the main governing party – Fine Gael – who lost significant support between 2011 and 2016 and we examine the vote choice of those voters who deserted the party. Hence, as reported in Table 4.4, we compare voters who voted for Fine Gael in 2011 and in 2017 and compare these ‘Fine Gael stickers’ to people who voted Fine Gael in 2011 but switched to either Fianna Fáil, Sinn Féin or Labour. Our rationale here is that there may be many reasons why people deserted Fine Gael, but once they do so, they must decide which of the other parties to support. While desertion may possibly be a function of disappointment with the governing performance of Fine Gael, it may be that post-desertion party choice is driven by assessing which party provides the closest socio-demographic and ideological match. Due to the low number of respondents, we simply present the results of a series of cross tabs, and only present statistically significant relationships (FG switchers to Labour, n = 70, and FG switchers to SF, n = 76). Switchers to Labour and to Sinn Féin are the most different in terms of social class and education, with almost all switchers to Labour being middle Table 4.4  Vote dynamics 2011 to 2016: Profiling FG stickers and switchers % 2011

2016

FG FG FG FG

FG FF Lab SF

%

Young* ABC1*** 32 33 39 47

60 53 86 41

%

%

Third level*

Weekly*

55 53 66 42

46 54 20 40

%

%

%

Rural** Unity*** Liberal** 37 34 14 36

27 33 44 67

38 39 72 53

Source: INES1 2016. *Statistically significant at 0.05, **0.01 and ***0.001 level. FG/FG, n = 770; FG/FF, n = 179; FG/Lab, n = 70, FG/SF, n = 76. ABC1 = middle-class demographic; FF = Fianna Fáil; FG = Fine Gael; Lab = Labour; SF = Sinn Féin.

Social and ideological bases of voting

79

class compared with only 41 per cent of switchers to Sinn Féin (and 66 per cent of switchers to Labour being third-level educated compared with only 42 per cent of switchers to Sinn Féin). The biggest ideological differences relate to being pro-unity and socially liberal. The pro-unity proportion of Fine Gael stickers is considerably lower than the pro-unity proportion of switchers to Sinn Féin. Similarly, the socially liberal proportion of Fine Gael stickers is considerably lower than the proportion of liberals among switchers to Labour.

Conclusion The rise of Sinn Féin as a major player in Irish politics has implications for how we understand the social and ideological bases of voting. The party’s strong socio-demographic profile (working class, left wing, and pro-unity) means that ‘political cleavage’-based models of vote choice that work well in many other democracies begin to work reasonably well in Ireland, at least in differentiating Sinn Féin voters. While social and ideological distinctions between Fianna Fáil, Fine Gael and Labour have traditionally been quite weak, and emerge again as quite weak in the analysis presented here, when one compares Sinn Féin voters to the voters of these three parties, clear differences emerge. Sinn Féin support is distinctly working class, pro-united Ireland and in favour of government action to reduce inequality. A similarly distinct profile of Sinn Féin supporters emerges when one focuses on those voters who switched to it from the main governing party, Fine Gael. Such switchers were significantly working class and pro-unity, in interesting contrast to Fine Gael-to-Labour switchers who were more likely to be middle class and socially liberal. So, the social and ideological distinctiveness of supporters of Irish parties is most noteworthy with respect to the role played by Sinn Féin in attracting a particular type of voter. What the analysis in this chapter shows is that we can shed light on the determinants of ideological voting in Ireland by taking the standard dimensions that have been used internationally – relating to economy, nation and sociomorality. While it has typically been hard in prior research to identify strong differences along these dimensions between the traditional parties of Fianna Fáil, Fine Gael and Labour, this analysis demonstrates that when the emerging big player – Sinn Féin – is compared with these older parties, then political cleavages and standard ideological dimensions can help us discern significant differences in party support bases. We hence build on the analysis of the previous election by Tilley and Garry which argued in favour of the emergence of ideology in the form of class politics, but we here highlight the distinct profile of Sinn Féin.

80

The post-crisis Irish voter

Notes 1 Due to the traditionally small proportion of non-Catholics in Ireland, the Catholic versus non-Catholic distinction has not often been focused on in survey-based analyses of Irish voting behaviour, with religiosity (how religious are you?) rather than religion (what religion are you?) used as an explanatory variable. However, Catholics now represent only 78 per cent of the Irish population, and so this is a substantial distinction in reality (http://www.cso.ie/en/media/csoie/releasespublications/documents/population/2017/ Chapter_8_Religion.pdf). In the survey data used in this chapter, Catholics are somewhat over-represented, at 85 per cent. 2 Many of our variables are likely to be related to each other. For example, people who are young also tend to be people who have low religious attendance. Regression modelling allows us to take this into account and identify whether it is primarily age or religiosity that is the driver of opinion on moral issues. In fact (Table 4.2, column 2), it turns out that age is no longer a significant predictor when all variables are taken into account, and social class and education level also emerge as non-significant.

References Evans, Geoffrey, and James Tilley. 2016. The New Class Politics in Britain: The Political Exclusion of the Working Class. Oxford: Oxford University Press. Evans, Geoffrey, Anthony Heath and Mansur Lalljee. 1996. ‘Measuring leftright and libertarian-authoritarian values in the British electorate’, The British Journal of Sociology. 47: 93–112. Garry, John. 2016. Consociation and Voting in Northern Ireland: Party Competition and Electoral Behaviour. Philadelphia: University of Pennsylvania Press. Garry, John. 2017. ‘Nationalist in the North and socialist in the South? Examining Sinn Féin’s support base on both sides of the border’, in Niall Ó Dochartaigh, Katy Hayward and Elizabeth Meehan (eds), Dynamics of Political Change in Ireland: Making and Breaking a Divided Island. London: Routledge, pp. 145–56. Garry, John, Fiachra Kennedy, Michael Marsh and Richard Sinnott. 2003. ‘What decided the election?’, in Michael Gallagher, Michael Marsh and Paul Mitchell (eds), How Ireland Voted 2002. London: Palgrave, pp. 119–42. Inglehart, Ronald, and Hans-Dieter Klingeman. 1976. ‘Party identification, ideological preference, and the left–right dimension among Western Mass Publics’, in Ian Budge, Ivor Crewe and Dennis Fairlie (eds), Party Identification and Beyond: Representations of Voting and Party Competition. New York: Wiley, pp. 243–73. Knutsen, Oddbjørn. 1997. ‘The partisan and value based component of leftright self-placement: A comparative study’, International Political Science Review. 18: 191–225. Laver, Michael. 1986. ‘Ireland: Politics with some social bases: An interpretation based on survey data’, Economic and Social Review. 17: 193–213.

Social and ideological bases of voting

81

Marsh, Michael. 2007. ‘Explanations of party choice’, in Michael Gallagher and Michael Marsh (eds), How Ireland Voted 2007: The Full Story of Ireland’s General Election. Basingstoke: Palgrave Macmillan, pp. 105–31. Marsh, Michael, and Kevin Cunningham. 2011. ‘A positive choice, or anyone but Fianna Fáil?’, in Michael Gallagher and Michael Marsh (eds), How Ireland Voted, 2011. Basingstoke: Palgrave Macmillan, pp. 172–204. Marsh, Michael, and Richard Sinnott. 1990. ‘How the voters decided’, in Michael Gallagher and Richard Sinnott (eds), How Ireland Voted: The Irish General Election, 1989. Galway: Galway University Press, pp. 68–93. Marsh, M., and R. Sinnott. 1993. ‘The Voters: stability and change’, in M. Gallagher and M. Laver (eds), How Ireland Voted, 1992. Dublin: PSAI Press, pp. 93–114. Marsh, M., and R. Sinnott. 1999. ‘The behaviour of the Irish voter’, in M. Marsh and P. Mitchell (eds), How Ireland Voted, 1997. Boulder: Westview Press in association with the PSAI Press, pp. 151–80. Marsh, Michael, Richard Sinnott, John Garry and Fiachra Kennedy. 2008. The Irish Voter: The Nature of Electoral Competition in the Republic of Ireland. Manchester: Manchester University Press. McElroy, Gail. 2017. ‘Party competition in Ireland: The emergence of a left-right dimension?’, in Michael Marsh, David M. Farrell and Gail. McElroy (eds), A Conservative Revolution? Electoral Change in Twenty-First Century Ireland. Oxford: Oxford University Press, pp. 11–27. Pulzer, Peter. 1967. Political Representation and Elections in Britain. London: Allen and Unwin. Sinnott, Richard. 1984. ‘Interpretations of the Irish party system’, European Journal of Political Research. 12: 289–307. Sinnott, Richard. 1995. Irish Voters Decide: Voting Behaviour in Elections and Referendums since 1918. Manchester: Manchester University Press. Tilley, James, and John Garry. 2017. ‘How economic catastrophe realigned Irish politics along economic divisions’, in Michael Marsh, David M. Farrell and Gail McElroy (eds), A Conservative Revolution? Electoral Change in TwentyFirst Century Ireland. Oxford: Oxford University Press, pp. 61–82. Whyte, John. 1974. ‘Ireland: Politics without social bases’, in R. Rose (ed.), Electoral Behaviour: A Comparative Handbook. New York: Free Press, pp. 619–51.

5 Party identification in the wake of the crisis: A nascent realignment? Rory Costello

Introduction Many commentators have sounded the death knell for party identification. For example, Dalton claims that we are witnessing a general process of partisan dealignment and that this trend ‘reflects long-term and enduring characteristics of advanced industrial societies’ (Dalton, 2002: 29). Like many other countries, Ireland experienced a sustained period of political dealignment, beginning in the 1970s (or earlier) and continuing right through to the new millennium. In Eurobarometer polls taken in the late 1970s, approximately two thirds of Irish respondents described themselves as being close to a political party; this had declined to 40 per cent by the mid-1990s (Mair and Marsh, 2004: 242). As reported below, just over one quarter of respondents admitted to feeling close to a party in Irish National Election Study (INES) surveys conducted in 2002 and 2007, and this fell even further in 2011. This is an important and, for many observers, worrying development. Partisanship is associated with political engagement, and is also seen by some as providing the stability necessary for a functioning representative democracy. As Rosenblum argues, ‘Partisans are carriers of a more extended story about the party than may be told by the candidates of the moment.’ Their long-term focus and attention to their party, even outside election years, acts as a ‘check on shortterm, arrant, political considerations’ by their party, as well as providing support and sustenance to the party following electoral defeat (Rosenblum, 2010: 355). A dealigned electorate, by contrast, is usually associated with disengagement, the growth of anti-establishment populism, and, above all, political instability (Green, Palmquist and Schickler, 2004: 222; Mair, 2013: 19). Dealignment is one potential explanation for the high levels of volatility observed in the 2011 and 2016 elections in Ireland, as argued by Marsh and McElroy (2016: 159). Not only were these the two most volatile elections in the history of the state, they also rank among the ten most volatile elections

Party identification in the wake of the crisis

83

in post-war Western Europe (Farrell and Suiter, 2016). Fianna Fáil, which had been the biggest party in every general election since 1927, lost over 70 per cent of its seats in 2011; Labour lost an even larger proportion of its seats in 2016. A number of new parties emerged during this period, and a record number of voters turned away from political parties altogether and voted instead for Independent candidates. If these are symptoms of a dealigned electorate, then electoral volatility is likely to be here to stay. Yet at other times and in other contexts, long-term declines in party identification have been reversed following a ‘critical’ election which disrupts the old political order and subsequent ‘cementing’ elections in which new political alignments become embedded (Miller and Shanks, 1996; Wattenberg, 1996: 138). The 2011 and 2016 Irish general elections could conceivably fit this pattern. As described by Burnham (1970), critical elections ‘are closely associated with abnormal stress in the socioeconomic system (and) are marked by ideological polarizations and issue-distances between the major parties which are exceptionally large by normal standards’. The 2011 election in Ireland was preceded by an unprecedented level of economic stress leading to the bailout by the EU and IMF, and the economic consensus that had dominated party politics during the 2000s gave way to greater political polarization. It is clear that the elections of 2011 and 2016 dramatically changed the Irish party system. The question that this chapter seeks to answer is whether the electoral turbulence was simply a symptom of a fundamentally dealigned electorate, or whether we are witnessing a realignment in Irish politics. In other words, has the number of floating voters increased in the wake of the crisis, or have people begun to form new party attachments that are likely to shape elections in the future? To address this question, this chapter will examine both the level and direction of party attachment in Ireland between 2002 and 2016. The chapter is organized as follows. First, the debates about the concept of party identification are introduced, and the meaning and measurement of party identification in Ireland are assessed using panel data from the Irish National Election Study (2002–2007). Next, arguments about partisan change (dealignment and realignment) are discussed, before presenting evidence on the evolution of partisanship in Ireland in the period 2002–2016, using the full set of INES studies. The chapter concludes by discussing what these findings imply for future elections in Ireland.

The meaning and measurement of party identification in Ireland Despite its importance to generations of electoral researchers, party identification remains a highly contested concept. The traditional account, found in ‘The American Voter’, is that party identification is a lasting psychological attachment to a party or (according to a more recent restatement) a sense of

84

The post-crisis Irish voter

belonging to a partisan group (Campbell et al., 1964: 67; Miller and Shanks, 1996: 120; Green, Palmquist and Schickler, 2004: 8). These attachments form relatively early in life, due to a combination of family and peer group influences and the political environment during an individual’s first experiences of voting (Miller and Shanks, 1996: 128–32). Once established, party attachments are expected to be highly stable, and become more ingrained over time. One of the main reasons for this stability, according to this view, is that partisanship is causally prior to and has a significant impact on many other political attitudes. In the words of Campbell et al. (1964: 76), ‘Identification with a party raises a perceptual screen through which the individual tends to see what is favourable to his partisan orientation.’ Partisanship is therefore seen as an exogenous factor exerting enormous influence on long-term electoral trends. There are a number of important critiques of this argument. A revisionist view assigns much less significance to party identification as a driver of other political attitudes. For Fiorina (1981: 84), partisanship is nothing more than a ‘running tally’ of an individual’s evaluations of party performance. This account treats partisanship as a rational assessment rather than as a psychological attachment. As such, it is much more susceptible to change in response to short-term political events, and not the important explanatory variable that the authors of ‘The American Voter’ assumed. Thomson (2017) provides evidence in support of this revisionist view of partisanship in the Irish case. A second line of criticism is that party identification has little meaning outside the United States. The usefulness of party identification as a concept is tied up with its stability: individuals may change their vote in response to short-term factors, but partisanship is far stickier, and hence a better predictor of future behaviour. However, in some European contexts, party identification was found to covary strongly with the vote (Thomassen, 1976; LeDuc, 1981). If party identification is synonymous with the vote, it loses its usefulness as an independent variable to explain long-term electoral patterns. The lack of applicability of the concept to many European countries is generally believed to derive from the very importance of political parties in structuring choice. Particularly in list-based electoral systems, where voters are essentially choosing between parties rather than candidates, voters are very unlikely to vote against their party identification. Furthermore, voters in many European countries have traditionally held strong class and religious identities, and these have been dominant over party identities. However, these arguments do not necessarily apply to the Irish case, which resembles the United States in a number of ways that makes party identification a more useful concept than it is in many other European countries. Voting in Ireland has traditionally not been driven by strong social group or class identities, so party identities could potentially fill this gap. The Irish electoral system is far more candidate-centred than most other European countries, so

Party identification in the wake of the crisis

85

how you vote and which party you support are not necessarily the same things. As Marsh notes, ‘If party identification proved to be a useful concept anywhere in Europe, Ireland would seem to be a prime candidate’ (Marsh, 2006: 491). Yet there are significant measurement issues that arise when applying the concept of party identification outside the United States, including in Ireland. The original formulation, still used in American National Election Studies, explicitly focuses on the respondent’s self-identity (‘Generally speaking, do you usually think of yourself as a Republican, a Democrat, an Independent or what?’). This approach has not been widely exported, for a number of reasons. On a practical level, it is difficult to formulate the question in this way in a multi-party context, particularly when there are no widely used nouns to refer to different partisan groups (Green, Palmquist and Schickler, 2004: 169).1 It is also more likely that individuals will identify with more than one party in a context where several parties are spread out across the ideological spectrum. It is more common in European surveys to measure party identification in terms of closeness to a party, and this is the approach used in the Irish National Election Study. The sequence of questions is as follows:2 1. ‘Do you usually think of yourself as close to any political party?’ [If yes] ‘Which party is that?’ 2. [If no] ‘Do you feel yourself a little closer to one of the political parties than the others?’ [If yes] ‘Which party is that’? ●●

●●

While this measure is arguably further removed from the concept of psychological attachment than the original formulation, it does focus the respondent’s attention on long-term attachment. Barnes et al. (1988) find that the two approaches produce scales that are highly correlated and similar in terms of reliability. Most previous research has measured partisanship in terms of the first of these two questions, treating those who reported feeling ‘a little closer’ to one of the parties as non-partisans (Marsh and Tilley, 2010; Thomson, 2017). However, movement over time from having no party attachment to feeling a little closer to one party (or vice versa) may be an important indicator of realignment or dealignment. This chapter therefore distinguishes between three types of voters: ‘partisans’ (those answering in affirmative to question 1 above); ‘leaners’ (those answering yes to the second question above); and ‘non-partisans’ (those answering no to both questions). Before proceeding to the main analysis in the next section, I first compare these three different categories of voters using the 2002–2007 INES panel data. The reason for this exercise is to determine whether ‘leaners’ are more like

86

The post-crisis Irish voter

partisans or non-partisans in terms of their voting behaviour and attitudes. This is an important consideration when it comes to interpreting changing levels of partisanship in the aftermath of the financial crisis. Partisan stability implies that party identifiers will be more stable in their vote over time compared to non-partisans. Table 5.1 compares the stability of voting preferences of our three categories of respondents in the period 2002 to 2007. Respondents are categorized according to whether they voted for the same party in 2002 and 2007 or changed their vote.3 The second column shows the percentage of respondents who voted for the same party in both elections. As expected, there are significant differences between those who identified with the party they voted for in 2002 and those who did not. For respondents who felt close to the party they voted for in 2002, 77 per cent went on to vote for the party again in 2007. For those categorized as ‘leaners’ in 2002, 64 per cent went on to vote for the party in 2007. In contrast, only 43 per cent of those who did not feel close to the party they voted for in 2002 went on to vote for the party again in 2007. A similar picture emerges when we look at respondents’ assessments in 2007 of the probability (on a scale of 1–10) that they will ever vote again for the party that they voted for in 2002. As reported in the right-hand column in Table 5.1, the average probability score is 8.63 for partisans, 7.88 for ‘leaners’ and 6.94 for non-partisans. In both of these sets of analyses, the difference between nonpartisans, ‘leaners’ and partisans are statistically significant. Both reported vote and probability to vote therefore tell the same story: partisans have more stable vote preferences over time than ‘leaners’, who in turn have more stable voting preferences than non-partisans.

Table 5.1  Stability of vote preference by party attachment, 2002–2007 Voted for same party in 2007 as 2002 (% respondents) 2002 Partisans 2002 Leaners 2002 Non-partisans All

77%** 64%** 43% 55%

2007 probability to vote for party voted for in 2002 (mean score) 8.63** 7.88** 6.94 7.54

Source: INES 2002, 2007. Note: **p